Dr. rer. nat. Stanislav Lange

• Discrete-Time Analysis of Multi-Component Queuing Networks under Renewal Approximation. Geißler, Stefan; Lange, Stanislav; Hasslinger, Gerhard; Tran-Gia, Phuoc; Hoßfeld, Tobias. In 34th International Teletraffic Congress (ITC). Shenzhen, China. 2023.
  • [ BibTeX ]
  • [ EndNote ]
  @inproceedings{info3-inproceedings-2023-1, author = {Geißler, Stefan and Lange, Stanislav and Hasslinger, Gerhard and Tran-Gia, Phuoc and Hoßfeld, Tobias}, booktitle = {34th International Teletraffic Congress (ITC). Shenzhen, China}, keywords = {sys:hidden:inproceedings}, month = 9, title = {Discrete-Time Analysis of Multi-Component Queuing Networks under Renewal Approximation}, year = 2023 }

  %0 Conference Paper %1 info3-inproceedings-2023-1 %A Geißler, Stefan %A Lange, Stanislav %A Hasslinger, Gerhard %A Tran-Gia, Phuoc %A Hoßfeld, Tobias %B 34th International Teletraffic Congress (ITC). Shenzhen, China %D 2023 %T Discrete-Time Analysis of Multi-Component Queuing Networks under Renewal Approximation

• Discrete-Time Analysis of Multi-Component Queuing Networks under Renewal Approximation. Geißler, Stefan; Lange, Stanislav; Haßllinger, Gerhard; Tran-Gia, Phuoc; Hoßfeld, Tobias. In 34st International Teletraffic Congress (ITC). , Shenzhen, China. 2022.
  • [ Abstract ]
  • [ BibTeX ]
  • [ EndNote ]
  The analytical and numerical performance evaluation of network components and distributed systems has been a staple in the networking community for many years. However, the ever growing complexity of modern systems and the need to gain detailed insights into systems consisting of many, interconnected components emphasizes the need for an extension to the classical single-component approach, and although approaches like Jackson Networks exist, their limited application scope lags behind the complexity of modern environments. To this end, we revisit existing models of the common Gi/Gi/1-oo queue, extend them to allow the concatenation of multiple queueing components, and evaluate the approximation error introduced through renewal approximation. We revisit previously performed parameter studies and evaluate the approximation error for a wide range of parameter combinations that we can solve through the power of modern computing equipment and efficient numerical implementations of our models. We show the main impact factors for the linear concatenation of queueing components as well as the split and superposition of processes. Our evaluations show that the renewal approximation can be applied to a wide range of parameters while still obtaining results within acceptable error margins.

  @inproceedings{info3-inproceedings-2022-17, abstract = {The analytical and numerical performance evaluation of network components and distributed systems has been a staple in the networking community for many years. However, the ever growing complexity of modern systems and the need to gain detailed insights into systems consisting of many, interconnected components emphasizes the need for an extension to the classical single-component approach, and although approaches like Jackson Networks exist, their limited application scope lags behind the complexity of modern environments. To this end, we revisit existing models of the common Gi/Gi/1-oo queue, extend them to allow the concatenation of multiple queueing components, and evaluate the approximation error introduced through renewal approximation. We revisit previously performed parameter studies and evaluate the approximation error for a wide range of parameter combinations that we can solve through the power of modern computing equipment and efficient numerical implementations of our models. We show the main impact factors for the linear concatenation of queueing components as well as the split and superposition of processes. Our evaluations show that the renewal approximation can be applied to a wide range of parameters while still obtaining results within acceptable error margins.}, author = {Geißler, Stefan and Lange, Stanislav and Haßllinger, Gerhard and Tran-Gia, Phuoc and Hoßfeld, Tobias}, booktitle = {34st International Teletraffic Congress (ITC). , Shenzhen, China}, keywords = {can}, month = 9, title = {Discrete-Time Analysis of Multi-Component Queuing Networks under Renewal Approximation}, year = 2022 }

  %0 Conference Paper %1 info3-inproceedings-2022-17 %A Geißler, Stefan %A Lange, Stanislav %A Haßllinger, Gerhard %A Tran-Gia, Phuoc %A Hoßfeld, Tobias %B 34st International Teletraffic Congress (ITC). , Shenzhen, China %D 2022 %T Discrete-Time Analysis of Multi-Component Queuing Networks under Renewal Approximation %X The analytical and numerical performance evaluation of network components and distributed systems has been a staple in the networking community for many years. However, the ever growing complexity of modern systems and the need to gain detailed insights into systems consisting of many, interconnected components emphasizes the need for an extension to the classical single-component approach, and although approaches like Jackson Networks exist, their limited application scope lags behind the complexity of modern environments. To this end, we revisit existing models of the common Gi/Gi/1-oo queue, extend them to allow the concatenation of multiple queueing components, and evaluate the approximation error introduced through renewal approximation. We revisit previously performed parameter studies and evaluate the approximation error for a wide range of parameter combinations that we can solve through the power of modern computing equipment and efficient numerical implementations of our models. We show the main impact factors for the linear concatenation of queueing components as well as the split and superposition of processes. Our evaluations show that the renewal approximation can be applied to a wide range of parameters while still obtaining results within acceptable error margins.

• Performance Evaluation of Hybrid Crowdsensing and Fixed Sensor Systems for Event Detection in Urban Environments. Hirth, Matthias; Seufert, Michael; Lange, Stanislav; Meixner, Markus; Tran-Gia, Phuoc. In Sensors, 21(17), pp. 5880 (1–19). 2021.
  • [ Abstract ]
  • [ BibTeX ]
  • [ EndNote ]
  Crowdsensing offers a cost-effective way to collect large amounts of environmental sensor data. However, the spatial distribution of crowdsensing sensors can hardly be influenced, as the participants carry the sensors, and, additionally, the quality of the crowdsensed data can vary significantly. Hybrid systems that use mobile users in conjunction with fixed sensors might help to overcome these limitations, as such system allow assessing the quality of the submitted crowdsensed data and provide sensor values where no crowdsensing data is available. In this work, we use a simulation study to first analyze a simple crowdsensing system concerning the detection performance of spatial events to highlight the potential and limitations of a pure crowdsourcing system. The results indicate that even if only a small share of inhabitants participates in crowdsensing, events that have locations correlated with the population density can be easily and quickly detected using such a system. On the contrary, events with uniformly randomly distributed locations are much harder to detect using a simple crowdsensing-based approach. A second evaluation shows that hybrid systems improve the detection probability and time. Finally, we illustrate how to compute the minimum number of fixed sensors for given detection time thresholds in our exemplary scenario.

  @article{info3-article-2021-4, abstract = {Crowdsensing offers a cost-effective way to collect large amounts of environmental sensor data. However, the spatial distribution of crowdsensing sensors can hardly be influenced, as the participants carry the sensors, and, additionally, the quality of the crowdsensed data can vary significantly. Hybrid systems that use mobile users in conjunction with fixed sensors might help to overcome these limitations, as such system allow assessing the quality of the submitted crowdsensed data and provide sensor values where no crowdsensing data is available. In this work, we use a simulation study to first analyze a simple crowdsensing system concerning the detection performance of spatial events to highlight the potential and limitations of a pure crowdsourcing system. The results indicate that even if only a small share of inhabitants participates in crowdsensing, events that have locations correlated with the population density can be easily and quickly detected using such a system. On the contrary, events with uniformly randomly distributed locations are much harder to detect using a simple crowdsensing-based approach. A second evaluation shows that hybrid systems improve the detection probability and time. Finally, we illustrate how to compute the minimum number of fixed sensors for given detection time thresholds in our exemplary scenario.}, author = {Hirth, Matthias and Seufert, Michael and Lange, Stanislav and Meixner, Markus and Tran-Gia, Phuoc}, journal = {Sensors}, keywords = {ucn}, month = 8, number = 17, pages = {5880 (1-19)}, title = {Performance Evaluation of Hybrid Crowdsensing and Fixed Sensor Systems for Event Detection in Urban Environments}, volume = 21, year = 2021 }

  %0 Journal Article %1 info3-article-2021-4 %A Hirth, Matthias %A Seufert, Michael %A Lange, Stanislav %A Meixner, Markus %A Tran-Gia, Phuoc %D 2021 %J Sensors %N 17 %P 5880 (1-19) %T Performance Evaluation of Hybrid Crowdsensing and Fixed Sensor Systems for Event Detection in Urban Environments %V 21 %X Crowdsensing offers a cost-effective way to collect large amounts of environmental sensor data. However, the spatial distribution of crowdsensing sensors can hardly be influenced, as the participants carry the sensors, and, additionally, the quality of the crowdsensed data can vary significantly. Hybrid systems that use mobile users in conjunction with fixed sensors might help to overcome these limitations, as such system allow assessing the quality of the submitted crowdsensed data and provide sensor values where no crowdsensing data is available. In this work, we use a simulation study to first analyze a simple crowdsensing system concerning the detection performance of spatial events to highlight the potential and limitations of a pure crowdsourcing system. The results indicate that even if only a small share of inhabitants participates in crowdsensing, events that have locations correlated with the population density can be easily and quickly detected using such a system. On the contrary, events with uniformly randomly distributed locations are much harder to detect using a simple crowdsensing-based approach. A second evaluation shows that hybrid systems improve the detection probability and time. Finally, we illustrate how to compute the minimum number of fixed sensors for given detection time thresholds in our exemplary scenario.
• Discrete-Time Modeling of NFV Accelerators that Exploit Batched Processing. Geissler, Stefan; Lange, Stanislav; Linguaglossa, Leonardo; Rossi, Dario; Zinner, Thomas; Hoßfeld, Tobias. In ACM Transactions on Modeling and Performance Evaluation of Computing Systems, 6(3), pp. 1–27. 2021.
  • [ BibTeX ]
  • [ EndNote ]
  @article{info3-article-2021-12, author = {Geissler, Stefan and Lange, Stanislav and Linguaglossa, Leonardo and Rossi, Dario and Zinner, Thomas and Hoßfeld, Tobias}, journal = {ACM Transactions on Modeling and Performance Evaluation of Computing Systems}, keywords = {can}, month = 9, number = 3, pages = {1-27}, title = {Discrete-Time Modeling of NFV Accelerators that Exploit Batched Processing}, volume = 6, year = 2021 }

  %0 Journal Article %1 info3-article-2021-12 %A Geissler, Stefan %A Lange, Stanislav %A Linguaglossa, Leonardo %A Rossi, Dario %A Zinner, Thomas %A Hoßfeld, Tobias %D 2021 %J ACM Transactions on Modeling and Performance Evaluation of Computing Systems %N 3 %P 1-27 %T Discrete-Time Modeling of NFV Accelerators that Exploit Batched Processing %V 6
• Discrete-time Analysis of Multicomponent GI/GI/1 Queueing Networks. Geissler, Stefan; Lange, Stanislav; Tran-Gia, Phuoc; Hoßfeld, Tobias. In Conference on Networked Systems (NetSys). 2021.
  • [ Abstract ]
  • [ BibTeX ]
  • [ EndNote ]
  In this work, we provide initial insights regarding the error introduced into multicomponent queueing systems by assuming the departure processes of arbitrary GI/GI/1-∞ queues to be renewal processes. To this end, we compute the sojourn time distribution as well as departure distributions of a linear chain of queueing components and compare the results to a simulation of the same system. By applying the renewal approximation, potential autocorrelations of the departure processes are lost. We investigate the magnitude of this error regarding both the sojourn time as well as interdeparture time distributions for a broad set of parameters. Although more indepth studies are needed, our results show that both distributions can be closely approximated, which allows the application of the model to asses the performance of real world NFV function chains.

  @inproceedings{info3-inproceedings-2021-4, abstract = {In this work, we provide initial insights regarding the error introduced into multicomponent queueing systems by assuming the departure processes of arbitrary GI/GI/1-∞ queues to be renewal processes. To this end, we compute the sojourn time distribution as well as departure distributions of a linear chain of queueing components and compare the results to a simulation of the same system. By applying the renewal approximation, potential autocorrelations of the departure processes are lost. We investigate the magnitude of this error regarding both the sojourn time as well as interdeparture time distributions for a broad set of parameters. Although more indepth studies are needed, our results show that both distributions can be closely approximated, which allows the application of the model to asses the performance of real world NFV function chains.}, author = {Geissler, Stefan and Lange, Stanislav and Tran-Gia, Phuoc and Hoßfeld, Tobias}, booktitle = {Conference on Networked Systems (NetSys)}, keywords = {sys:hidden:inproceedings}, month = 11, title = {Discrete-time Analysis of Multicomponent GI/GI/1 Queueing Networks}, year = 2021 }

  %0 Conference Paper %1 info3-inproceedings-2021-4 %A Geissler, Stefan %A Lange, Stanislav %A Tran-Gia, Phuoc %A Hoßfeld, Tobias %B Conference on Networked Systems (NetSys) %D 2021 %T Discrete-time Analysis of Multicomponent GI/GI/1 Queueing Networks %X In this work, we provide initial insights regarding the error introduced into multicomponent queueing systems by assuming the departure processes of arbitrary GI/GI/1-∞ queues to be renewal processes. To this end, we compute the sojourn time distribution as well as departure distributions of a linear chain of queueing components and compare the results to a simulation of the same system. By applying the renewal approximation, potential autocorrelations of the departure processes are lost. We investigate the magnitude of this error regarding both the sojourn time as well as interdeparture time distributions for a broad set of parameters. Although more indepth studies are needed, our results show that both distributions can be closely approximated, which allows the application of the model to asses the performance of real world NFV function chains.

• Survey of Performance Acceleration Techniques for Network Function Virtualization. Linguaglossa, Leonardo; Lange, Stanislav; Pontarelli, Salvatore; Retvari, Gabor; Rossi, Dario; Zinner, Thomas; Bificulo, Roberto; Jarschel, Michael; Bianchi, Giuseppe. In Proceedings of the IEEE. 2019.
  • [ Abstract ]
  • [ BibTeX ]
  • [ EndNote ]
  The ongoing network softwarization trend holds the promise to revolutionize network infrastructures by making them more flexible, reconfigurable, portable, and more adaptive than ever. Still, the migration from hard-coded/hardwired network functions towards their software-programmable counterparts comes along with the need for tailored optimizations and acceleration techniques, so as to avoid, or at least mitigate, the throughput/latency performance degradation with respect to fixed function network elements. The contribution of this article is twofold. First, we provide a comprehensive overview of the host-based Network Function Virtualization (NFV) ecosystem, covering a broad range of techniques, from low level hardware acceleration and bump-in-the-wire offloading approaches, to high-level software acceleration solutions, including the virtualization technique itself. Second, we derive guidelines regarding the design, development, and operation of NFV-based deployments that meet the flexibility and scalability requirements of modern communication networks.

  @article{info3-article-2019-1, abstract = {The ongoing network softwarization trend holds the promise to revolutionize network infrastructures by making them more flexible, reconfigurable, portable, and more adaptive than ever. Still, the migration from hard-coded/hardwired network functions towards their software-programmable counterparts comes along with the need for tailored optimizations and acceleration techniques, so as to avoid, or at least mitigate, the throughput/latency performance degradation with respect to fixed function network elements. The contribution of this article is twofold. First, we provide a comprehensive overview of the host-based Network Function Virtualization (NFV) ecosystem, covering a broad range of techniques, from low level hardware acceleration and bump-in-the-wire offloading approaches, to high-level software acceleration solutions, including the virtualization technique itself. Second, we derive guidelines regarding the design, development, and operation of NFV-based deployments that meet the flexibility and scalability requirements of modern communication networks.}, author = {Linguaglossa, Leonardo and Lange, Stanislav and Pontarelli, Salvatore and Retvari, Gabor and Rossi, Dario and Zinner, Thomas and Bificulo, Roberto and Jarschel, Michael and Bianchi, Giuseppe}, journal = {Proceedings of the IEEE}, keywords = {sys:hidden:article}, month = 1, title = {Survey of Performance Acceleration Techniques for Network Function Virtualization}, year = 2019 }

  %0 Journal Article %1 info3-article-2019-1 %A Linguaglossa, Leonardo %A Lange, Stanislav %A Pontarelli, Salvatore %A Retvari, Gabor %A Rossi, Dario %A Zinner, Thomas %A Bificulo, Roberto %A Jarschel, Michael %A Bianchi, Giuseppe %D 2019 %J Proceedings of the IEEE %T Survey of Performance Acceleration Techniques for Network Function Virtualization %X The ongoing network softwarization trend holds the promise to revolutionize network infrastructures by making them more flexible, reconfigurable, portable, and more adaptive than ever. Still, the migration from hard-coded/hardwired network functions towards their software-programmable counterparts comes along with the need for tailored optimizations and acceleration techniques, so as to avoid, or at least mitigate, the throughput/latency performance degradation with respect to fixed function network elements. The contribution of this article is twofold. First, we provide a comprehensive overview of the host-based Network Function Virtualization (NFV) ecosystem, covering a broad range of techniques, from low level hardware acceleration and bump-in-the-wire offloading approaches, to high-level software acceleration solutions, including the virtualization technique itself. Second, we derive guidelines regarding the design, development, and operation of NFV-based deployments that meet the flexibility and scalability requirements of modern communication networks.
• Discrete-Time Modeling of NFV Accelerators that Exploit Batched Processing. Lange, Stanislav; Linguaglossa, Leonardo; Geissler, Stefan; Rossi, Dario; Zinner, Thomas. In IEEE Conference on Computer Communications (Infocom). 2019.
  • [ BibTeX ]
  • [ EndNote ]
  • [ Download ]
  @inproceedings{info3-inproceedings-2019-4, author = {Lange, Stanislav and Linguaglossa, Leonardo and Geissler, Stefan and Rossi, Dario and Zinner, Thomas}, booktitle = {IEEE Conference on Computer Communications (Infocom)}, keywords = {sendate}, month = 4, title = {Discrete-Time Modeling of NFV Accelerators that Exploit Batched Processing}, year = 2019 }

  %0 Conference Paper %1 info3-inproceedings-2019-4 %A Lange, Stanislav %A Linguaglossa, Leonardo %A Geissler, Stefan %A Rossi, Dario %A Zinner, Thomas %B IEEE Conference on Computer Communications (Infocom) %D 2019 %T Discrete-Time Modeling of NFV Accelerators that Exploit Batched Processing
• KOMon - Kernel-based Online Monitoring of VNF Packet Processing Times. Geissler, Stefan; Lange, Stanislav; Wamser, Florian; Zinner, Thomas; Hoßfeld, Tobias. In 2019 International Conference on Networked Systems (NetSys), Best Paper Award. 2019.
  • [ BibTeX ]
  • [ EndNote ]
  • [ Download ]
  @inproceedings{info3-inproceedings-2019-5, author = {Geissler, Stefan and Lange, Stanislav and Wamser, Florian and Zinner, Thomas and Hoßfeld, Tobias}, booktitle = {2019 International Conference on Networked Systems (NetSys), Best Paper Award}, keywords = {sendate}, month = 3, title = {KOMon - Kernel-based Online Monitoring of VNF Packet Processing Times}, year = 2019 }

  %0 Conference Paper %1 info3-inproceedings-2019-5 %A Geissler, Stefan %A Lange, Stanislav %A Wamser, Florian %A Zinner, Thomas %A Hoßfeld, Tobias %B 2019 International Conference on Networked Systems (NetSys), Best Paper Award %D 2019 %T KOMon - Kernel-based Online Monitoring of VNF Packet Processing Times
• Discrete-Time Analysis of the Blockchain Distributed Ledger Technology. Geissler, Stefan; Prantl, Thomas; Lange, Stanislav; Wamser, Florian; Hoßfeld, Tobias. In 31st International Teletraffic Congress (ITC). , Budapest, Hungary. 2019.
  • [ Abstract ]
  • [ BibTeX ]
  • [ EndNote ]
  • [ Download ]
  Blockchain and distributed ledger technologies have become more and more popular and widespread during recent years. After the initial hype about the technology and many cryptocurrency related use cases, the technology slowly starts to make its way into other domains like food tracking and document management. In order to further contribute to the search of what this technology can be used for, more detailed performance evaluations are required in order to investigate key performance indicators and general limits of the technology. To this end, we develop a discrete-time queueing model that allows a detailed evaluation of the characteristics of a blockchain system, such as the transaction waiting time distribution. Furthermore, we validate the model by comparing the results to values obtained from measurements in a closed lab environment based on the Ethereum blockchain.

  @inproceedings{info3-inproceedings-2019-19, abstract = {Blockchain and distributed ledger technologies have become more and more popular and widespread during recent years. After the initial hype about the technology and many cryptocurrency related use cases, the technology slowly starts to make its way into other domains like food tracking and document management. In order to further contribute to the search of what this technology can be used for, more detailed performance evaluations are required in order to investigate key performance indicators and general limits of the technology. To this end, we develop a discrete-time queueing model that allows a detailed evaluation of the characteristics of a blockchain system, such as the transaction waiting time distribution. Furthermore, we validate the model by comparing the results to values obtained from measurements in a closed lab environment based on the Ethereum blockchain.}, author = {Geissler, Stefan and Prantl, Thomas and Lange, Stanislav and Wamser, Florian and Hoßfeld, Tobias}, booktitle = {31st International Teletraffic Congress (ITC). , Budapest, Hungary}, keywords = {net-man}, month = 8, title = {Discrete-Time Analysis of the Blockchain Distributed Ledger Technology}, year = 2019 }

  %0 Conference Paper %1 info3-inproceedings-2019-19 %A Geissler, Stefan %A Prantl, Thomas %A Lange, Stanislav %A Wamser, Florian %A Hoßfeld, Tobias %B 31st International Teletraffic Congress (ITC). , Budapest, Hungary %D 2019 %T Discrete-Time Analysis of the Blockchain Distributed Ledger Technology %X Blockchain and distributed ledger technologies have become more and more popular and widespread during recent years. After the initial hype about the technology and many cryptocurrency related use cases, the technology slowly starts to make its way into other domains like food tracking and document management. In order to further contribute to the search of what this technology can be used for, more detailed performance evaluations are required in order to investigate key performance indicators and general limits of the technology. To this end, we develop a discrete-time queueing model that allows a detailed evaluation of the characteristics of a blockchain system, such as the transaction waiting time distribution. Furthermore, we validate the model by comparing the results to values obtained from measurements in a closed lab environment based on the Ethereum blockchain.

• Optimization of Controller Placement and Information Flow in Softwarized Networks. Technical Report (PhD dissertation), . Lange, Stanislav. PhD dissertation. Universitiy of Würzburg, 2018, December.
  • [ BibTeX ]
  • [ EndNote ]
  • [ Download ]
  @phdthesis{info3-dissertation-2018-3, author = {Lange, Stanislav}, keywords = {sys:hidden:supervisor:phuoc.tran-gia}, month = 12, school = {Universitiy of Würzburg}, title = {Optimization of Controller Placement and Information Flow in Softwarized Networks}, year = 2018 }

  %0 Thesis %1 info3-dissertation-2018-3 %A Lange, Stanislav %D 2018 %T Optimization of Controller Placement and Information Flow in Softwarized Networks
• Deriving YouTube Playout Phases from Encrypted Packet Level Traffic. Geissler, Stefan; Lange, Stanislav; Wamser, Florian; Hoßfeld, Tobias. 2018.
  • [ Abstract ]
  • [ BibTeX ]
  • [ EndNote ]
  From the point of view of telecommunication providers, video streaming is one of the most demanding applications in today's Internet. Over 73\% of the total global network traffic has been attributed to video streaming applications in 2016. In this work, we provide a first step towards a better understanding of the packet level behavior of video streaming traffic to enable more efficient traffic engineering in the future. We perform a measurement study with the popular video streaming platform YouTube and show that the different playout phases of a video streaming session can not only be observed by evaluating application layer metrics, but also from raw and encrypted packet level traces.

  @misc{info3-poster-2018-1, abstract = {From the point of view of telecommunication providers, video streaming is one of the most demanding applications in today's Internet. Over 73\% of the total global network traffic has been attributed to video streaming applications in 2016. In this work, we provide a first step towards a better understanding of the packet level behavior of video streaming traffic to enable more efficient traffic engineering in the future. We perform a measurement study with the popular video streaming platform YouTube and show that the different playout phases of a video streaming session can not only be observed by evaluating application layer metrics, but also from raw and encrypted packet level traces.}, author = {Geissler, Stefan and Lange, Stanislav and Wamser, Florian and Hoßfeld, Tobias}, booktitle = {30th International Teletraffic Congress (ITC 30)}, keywords = {sys:hidden:poster}, title = {Deriving YouTube Playout Phases from Encrypted Packet Level Traffic}, year = 2018 }

  %0 Generic %1 info3-poster-2018-1 %A Geissler, Stefan %A Lange, Stanislav %A Wamser, Florian %A Hoßfeld, Tobias %B 30th International Teletraffic Congress (ITC 30) %D 2018 %T Deriving YouTube Playout Phases from Encrypted Packet Level Traffic %X From the point of view of telecommunication providers, video streaming is one of the most demanding applications in today's Internet. Over 73\% of the total global network traffic has been attributed to video streaming applications in 2016. In this work, we provide a first step towards a better understanding of the packet level behavior of video streaming traffic to enable more efficient traffic engineering in the future. We perform a measurement study with the popular video streaming platform YouTube and show that the different playout phases of a video streaming session can not only be observed by evaluating application layer metrics, but also from raw and encrypted packet level traces.
• Evaluation of a Distributed Control Plane for Managing Heterogeneous SDN-enabled and Legacy Networks. Gray, Nicholas; Lange, Stanislav; Zinner, Thomas; Pfaff, Benedikt; Hock, David. In 7th International Conference on Communications and Electronics (ICCE). Hue, Vietnam, 2018.
  • [ Abstract ]
  • [ BibTeX ]
  • [ EndNote ]
  • [ Download ]
  With the increasing number of devices, protocols and applications, today’s networks are becoming more and more complex. Hence, Software-defined Networking (SDN) tries to ad- dress this issue by separating the data from the control plane and by providing centralized interfaces for network configuration. As legacy devices cannot be replaced instantly due to high costs, both network segments have to be operated in coexistence with defined joints at their edges. To ensure a smooth operation, both controlling instances of these segments are required to exchange information. In this work, we design and implement a data model for storing the information needed to keep the controller and a Network Management System (NMS) synchronized, which are responsible for configuring the SDN-enabled network and the legacy devices respectively. For this, we implement and evaluate a total of three different synchronization strategies by the example of an SDN-based Bring Your Own Device (BYOD) use case. Index Terms—SDN, NMS, Management, Integration, Hetero- geneous Networks, Distributed Control Plane.

  @inproceedings{info3-inproceedings-2018-33, abstract = {With the increasing number of devices, protocols and applications, today’s networks are becoming more and more complex. Hence, Software-defined Networking (SDN) tries to ad- dress this issue by separating the data from the control plane and by providing centralized interfaces for network configuration. As legacy devices cannot be replaced instantly due to high costs, both network segments have to be operated in coexistence with defined joints at their edges. To ensure a smooth operation, both controlling instances of these segments are required to exchange information. In this work, we design and implement a data model for storing the information needed to keep the controller and a Network Management System (NMS) synchronized, which are responsible for configuring the SDN-enabled network and the legacy devices respectively. For this, we implement and evaluate a total of three different synchronization strategies by the example of an SDN-based Bring Your Own Device (BYOD) use case. Index Terms—SDN, NMS, Management, Integration, Hetero- geneous Networks, Distributed Control Plane.}, address = {Hue, Vietnam}, author = {Gray, Nicholas and Lange, Stanislav and Zinner, Thomas and Pfaff, Benedikt and Hock, David}, booktitle = {7th International Conference on Communications and Electronics (ICCE)}, keywords = {sardine}, month = 7, title = {Evaluation of a Distributed Control Plane for Managing Heterogeneous SDN-enabled and Legacy Networks}, year = 2018 }

  %0 Conference Paper %1 info3-inproceedings-2018-33 %A Gray, Nicholas %A Lange, Stanislav %A Zinner, Thomas %A Pfaff, Benedikt %A Hock, David %B 7th International Conference on Communications and Electronics (ICCE) %C Hue, Vietnam %D 2018 %T Evaluation of a Distributed Control Plane for Managing Heterogeneous SDN-enabled and Legacy Networks %X With the increasing number of devices, protocols and applications, today’s networks are becoming more and more complex. Hence, Software-defined Networking (SDN) tries to ad- dress this issue by separating the data from the control plane and by providing centralized interfaces for network configuration. As legacy devices cannot be replaced instantly due to high costs, both network segments have to be operated in coexistence with defined joints at their edges. To ensure a smooth operation, both controlling instances of these segments are required to exchange information. In this work, we design and implement a data model for storing the information needed to keep the controller and a Network Management System (NMS) synchronized, which are responsible for configuring the SDN-enabled network and the legacy devices respectively. For this, we implement and evaluate a total of three different synchronization strategies by the example of an SDN-based Bring Your Own Device (BYOD) use case. Index Terms—SDN, NMS, Management, Integration, Hetero- geneous Networks, Distributed Control Plane.
• Identification of Delay Thresholds Representing the Perceived Quality of Enterprise Applications. Borchert, Kathrin; Lange, Stanislav; Zinner, Thomas; Hirth, Matthias. In The 2nd International Workshop on Quality of Experience Management. Sardinia, Italy, 2018.
  • [ BibTeX ]
  • [ EndNote ]
  @inproceedings{info3-inproceedings-2018-11, address = {Sardinia, Italy}, author = {Borchert, Kathrin and Lange, Stanislav and Zinner, Thomas and Hirth, Matthias}, booktitle = {The 2nd International Workshop on Quality of Experience Management}, keywords = {ua-net}, month = 5, title = {Identification of Delay Thresholds Representing the Perceived Quality of Enterprise Applications}, year = 2018 }

  %0 Conference Paper %1 info3-inproceedings-2018-11 %A Borchert, Kathrin %A Lange, Stanislav %A Zinner, Thomas %A Hirth, Matthias %B The 2nd International Workshop on Quality of Experience Management %C Sardinia, Italy %D 2018 %T Identification of Delay Thresholds Representing the Perceived Quality of Enterprise Applications
• Integrating Network Management Information into the SDN Control Plane. Lange, Stanislav; Reinhart, Lorenz; Zinner, Thomas; Hock, David; Gray, Nicholas; Tran-Gia, Phuoc. In IEEE/IFIP Network Operations and Management Symposium (NOMS). 2018.
  • [ Abstract ]
  • [ BibTeX ]
  • [ EndNote ]
  With software defined networking (SDN), operators benefit from a higher flexibility, cost efficiency, as well as programmability of their networks. Since modern networks are comprised of a multitude of heterogeneous devices and also include non-SDN legacy devices, network management systems (NMSs) are often used in order to monitor and configure the network. Although both, the SDN controller and the NMS, have a centralized view of the network, they operate at different time scales and deal with information at different levels of granularity. In this work, we investigate the impact on the network performance when an NMS regularly provides information to an SDN controller. To this end, we design, implement, and compare three interaction mechanisms based on the ONOS controller. These represent different trade-offs regarding the complexity of the resulting system and its performance. In addition to the default ONOS controller, we develop two extended versions. One performs hash-based load balancing on equal cost paths while the other utilizes external NMS information via ONOS's intent and annotation framework to optimize control plane decisions. In addition to evaluations that show a significant performance improvement when using the optimized controllers, we present a parameter study that highlights the performance impact of network characteristics like the flow interarrival time, the flow duration, and the number of active flows.

  @inproceedings{info3-inproceedings-2018-10, abstract = {With software defined networking (SDN), operators benefit from a higher flexibility, cost efficiency, as well as programmability of their networks. Since modern networks are comprised of a multitude of heterogeneous devices and also include non-SDN legacy devices, network management systems (NMSs) are often used in order to monitor and configure the network. Although both, the SDN controller and the NMS, have a centralized view of the network, they operate at different time scales and deal with information at different levels of granularity. In this work, we investigate the impact on the network performance when an NMS regularly provides information to an SDN controller. To this end, we design, implement, and compare three interaction mechanisms based on the ONOS controller. These represent different trade-offs regarding the complexity of the resulting system and its performance. In addition to the default ONOS controller, we develop two extended versions. One performs hash-based load balancing on equal cost paths while the other utilizes external NMS information via ONOS's intent and annotation framework to optimize control plane decisions. In addition to evaluations that show a significant performance improvement when using the optimized controllers, we present a parameter study that highlights the performance impact of network characteristics like the flow interarrival time, the flow duration, and the number of active flows.}, author = {Lange, Stanislav and Reinhart, Lorenz and Zinner, Thomas and Hock, David and Gray, Nicholas and Tran-Gia, Phuoc}, booktitle = {IEEE/IFIP Network Operations and Management Symposium (NOMS)}, keywords = {sdnperf}, month = 4, title = {Integrating Network Management Information into the SDN Control Plane}, year = 2018 }

  %0 Conference Paper %1 info3-inproceedings-2018-10 %A Lange, Stanislav %A Reinhart, Lorenz %A Zinner, Thomas %A Hock, David %A Gray, Nicholas %A Tran-Gia, Phuoc %B IEEE/IFIP Network Operations and Management Symposium (NOMS) %D 2018 %T Integrating Network Management Information into the SDN Control Plane %X With software defined networking (SDN), operators benefit from a higher flexibility, cost efficiency, as well as programmability of their networks. Since modern networks are comprised of a multitude of heterogeneous devices and also include non-SDN legacy devices, network management systems (NMSs) are often used in order to monitor and configure the network. Although both, the SDN controller and the NMS, have a centralized view of the network, they operate at different time scales and deal with information at different levels of granularity. In this work, we investigate the impact on the network performance when an NMS regularly provides information to an SDN controller. To this end, we design, implement, and compare three interaction mechanisms based on the ONOS controller. These represent different trade-offs regarding the complexity of the resulting system and its performance. In addition to the default ONOS controller, we develop two extended versions. One performs hash-based load balancing on equal cost paths while the other utilizes external NMS information via ONOS's intent and annotation framework to optimize control plane decisions. In addition to evaluations that show a significant performance improvement when using the optimized controllers, we present a parameter study that highlights the performance impact of network characteristics like the flow interarrival time, the flow duration, and the number of active flows.
• Estimating the Flow Rule Installation Time of SDN Switches when Facing Control Plane Delay. Nguyen-Ngoc, Anh; Lange, Stanislav; Geissler, Stefan; Zinner, Thomas; Tran-Gia, Phuoc. In 19th International GI/ITG Conference on “Measurement, Modelling and Evaluation of Computing Systems”. Erlangen, 2018.
  • [ Abstract ]
  • [ BibTeX ]
  • [ EndNote ]
  • [ Download ]
  The software defined networking (SDN) paradigm has numerous benefits for network operators, including cost aspects, flexibility, and programmability. In the centralized SDN architecture, the controller can order the installation of flow rules in the switches it manages via FlowMod messages. Since the processing time of these messages has a direct impact on the reaction time of the network, it is a key performance indicator for switches and quantifying it in a reliable manner is required for ensuring state consistency between the control and the data plane. Furthermore, real world deployments not only consist of different data plane hardware, but may feature varying control plane delays. Hence, in this work, we investigate the impact of such a delay on the FlowMod processing time of OpenFlow switches. Firstly, we identify a significant heterogeneity between data plane hardware in terms of processing times as well as the underlying TCP-level behavior. Secondly, we show that despite this heterogeneity, combining switch specific information with delay measurements at the controller can be used to reliably infer FlowMod processing times. We confirm our results with measurements in a dedicated testbed that is comprised of three different hardware switches, three different SDN controllers, and several high precision measurement devices.

  @inproceedings{info3-inproceedings-2018-4, abstract = {The software defined networking (SDN) paradigm has numerous benefits for network operators, including cost aspects, flexibility, and programmability. In the centralized SDN architecture, the controller can order the installation of flow rules in the switches it manages via FlowMod messages. Since the processing time of these messages has a direct impact on the reaction time of the network, it is a key performance indicator for switches and quantifying it in a reliable manner is required for ensuring state consistency between the control and the data plane. Furthermore, real world deployments not only consist of different data plane hardware, but may feature varying control plane delays. Hence, in this work, we investigate the impact of such a delay on the FlowMod processing time of OpenFlow switches. Firstly, we identify a significant heterogeneity between data plane hardware in terms of processing times as well as the underlying TCP-level behavior. Secondly, we show that despite this heterogeneity, combining switch specific information with delay measurements at the controller can be used to reliably infer FlowMod processing times. We confirm our results with measurements in a dedicated testbed that is comprised of three different hardware switches, three different SDN controllers, and several high precision measurement devices.}, address = {Erlangen}, author = {Nguyen-Ngoc, Anh and Lange, Stanislav and Geissler, Stefan and Zinner, Thomas and Tran-Gia, Phuoc}, booktitle = {19th International GI/ITG Conference on “Measurement, Modelling and Evaluation of Computing Systems”}, keywords = {sendate}, month = 2, title = {Estimating the Flow Rule Installation Time of SDN Switches when Facing Control Plane Delay}, year = 2018 }

  %0 Conference Paper %1 info3-inproceedings-2018-4 %A Nguyen-Ngoc, Anh %A Lange, Stanislav %A Geissler, Stefan %A Zinner, Thomas %A Tran-Gia, Phuoc %B 19th International GI/ITG Conference on “Measurement, Modelling and Evaluation of Computing Systems” %C Erlangen %D 2018 %T Estimating the Flow Rule Installation Time of SDN Switches when Facing Control Plane Delay %X The software defined networking (SDN) paradigm has numerous benefits for network operators, including cost aspects, flexibility, and programmability. In the centralized SDN architecture, the controller can order the installation of flow rules in the switches it manages via FlowMod messages. Since the processing time of these messages has a direct impact on the reaction time of the network, it is a key performance indicator for switches and quantifying it in a reliable manner is required for ensuring state consistency between the control and the data plane. Furthermore, real world deployments not only consist of different data plane hardware, but may feature varying control plane delays. Hence, in this work, we investigate the impact of such a delay on the FlowMod processing time of OpenFlow switches. Firstly, we identify a significant heterogeneity between data plane hardware in terms of processing times as well as the underlying TCP-level behavior. Secondly, we show that despite this heterogeneity, combining switch specific information with delay measurements at the controller can be used to reliably infer FlowMod processing times. We confirm our results with measurements in a dedicated testbed that is comprised of three different hardware switches, three different SDN controllers, and several high precision measurement devices.
• Performance Benchmarking of Network Function Chain Placement Algorithms. Grigorjew, Alexej; Lange, Stanislav; Zinner, Thomas; Tran-Gia, Phuoc. In 19th International GI/ITG Conference on Measurement, Modelling and Evaluation of Computing Systems (MMB). 2018.
  • [ BibTeX ]
  • [ EndNote ]
  • [ Download ]
  @inproceedings{info3-inproceedings-2018-1, author = {Grigorjew, Alexej and Lange, Stanislav and Zinner, Thomas and Tran-Gia, Phuoc}, booktitle = {19th International GI/ITG Conference on Measurement, Modelling and Evaluation of Computing Systems (MMB)}, keywords = {sendate}, month = 2, title = {Performance Benchmarking of Network Function Chain Placement Algorithms}, year = 2018 }

  %0 Conference Paper %1 info3-inproceedings-2018-1 %A Grigorjew, Alexej %A Lange, Stanislav %A Zinner, Thomas %A Tran-Gia, Phuoc %B 19th International GI/ITG Conference on Measurement, Modelling and Evaluation of Computing Systems (MMB) %D 2018 %T Performance Benchmarking of Network Function Chain Placement Algorithms
• Performance Evaluation of Mobile Crowdsensing for Event Detection. Hirth, Matthias; Lange, Stanislav; Seufert, Michael; Tran-Gia, Phuoc. In Proceedings of the 5th International Workshop on Crowd Assisted Sensing, Pervasive Systems and Communications. Athens, Greece, 2018.
  • [ Abstract ]
  • [ BibTeX ]
  • [ EndNote ]
  Crowdsensing offers a cost effective way to collect large amounts of sensor data. However, in contrast to fixed sensor deployments, the spatial distribution of the sensors can hardly be influence, as the sensors are carried by participants of the crowdsensing system. This in turn raises the question about the performance of such systems with respect to the detection probability and detection time of spatial events. In order to address this question, we analyze the performance of such a crowdsensing system by means of simulation. We use the traffic infrastructure of a small size city in Germany and simulate the inhabitants’ movement patterns with the well established SUMO mobility generator. Our results show that even if only a small share of inhabitants participates in crowdsensing, events, which have locations that are correlated with the population density, can be easily and quickly detected using such a system. On the contrary, events whose locations are uniformly randomly distributed are much harder to detect using a crowdsensing approach.

  @inproceedings{info3-inproceedings-2017-21, abstract = {Crowdsensing offers a cost effective way to collect large amounts of sensor data. However, in contrast to fixed sensor deployments, the spatial distribution of the sensors can hardly be influence, as the sensors are carried by participants of the crowdsensing system. This in turn raises the question about the performance of such systems with respect to the detection probability and detection time of spatial events. In order to address this question, we analyze the performance of such a crowdsensing system by means of simulation. We use the traffic infrastructure of a small size city in Germany and simulate the inhabitants’ movement patterns with the well established SUMO mobility generator. Our results show that even if only a small share of inhabitants participates in crowdsensing, events, which have locations that are correlated with the population density, can be easily and quickly detected using such a system. On the contrary, events whose locations are uniformly randomly distributed are much harder to detect using a crowdsensing approach.}, address = {Athens, Greece}, author = {Hirth, Matthias and Lange, Stanislav and Seufert, Michael and Tran-Gia, Phuoc}, booktitle = {Proceedings of the 5th International Workshop on Crowd Assisted Sensing, Pervasive Systems and Communications}, keywords = {sys:hidden:inproceedings}, month = 3, title = {Performance Evaluation of Mobile Crowdsensing for Event Detection}, year = 2018 }

  %0 Conference Paper %1 info3-inproceedings-2017-21 %A Hirth, Matthias %A Lange, Stanislav %A Seufert, Michael %A Tran-Gia, Phuoc %B Proceedings of the 5th International Workshop on Crowd Assisted Sensing, Pervasive Systems and Communications %C Athens, Greece %D 2018 %T Performance Evaluation of Mobile Crowdsensing for Event Detection %X Crowdsensing offers a cost effective way to collect large amounts of sensor data. However, in contrast to fixed sensor deployments, the spatial distribution of the sensors can hardly be influence, as the sensors are carried by participants of the crowdsensing system. This in turn raises the question about the performance of such systems with respect to the detection probability and detection time of spatial events. In order to address this question, we analyze the performance of such a crowdsensing system by means of simulation. We use the traffic infrastructure of a small size city in Germany and simulate the inhabitants’ movement patterns with the well established SUMO mobility generator. Our results show that even if only a small share of inhabitants participates in crowdsensing, events, which have locations that are correlated with the population density, can be easily and quickly detected using such a system. On the contrary, events whose locations are uniformly randomly distributed are much harder to detect using a crowdsensing approach.
• Benchmarking the ONOS Controller with OFCProbe. Nguyen-Ngoc, Anh; Raffeck, Simon; Lange, Stanislav; Geissler, Stefan; Zinner, Thomas; Tran-Gia, Phuoc. In 7th International Conference on Communications and Electronics (ICCE). Hue, Vietnam, 2018.
  • [ Abstract ]
  • [ BibTeX ]
  • [ EndNote ]
  The increased popularity of Software Defined Networking (SDN) drives the development and research of all involved components. In particular, state-of-the-art SDN controllers are a central part of the resulting network architecture and have a large impact on the performance of the entire network. Hence, it is necessary to quantify various performance indicators of the SDN controller in order to assess its feasibility in the context of different network characteristics such as topology and network size w.r.t. the number of nodes and links. To this end, we extend the open-source controller benchmarking tool OFCProbe with several novel modules that are capable of recording relevant performance metrics and demonstrate their use in the context of benchmarking the ONOS controller. Furthermore, we investigate the impact of different types of topology characteristics on the performance of the controller. In this work, we focus on several performance indicators for SDN controllers. First, the time that the ONOS controller needs to discover the network topology and update its internal representation of the topology. Second, we investigate the duration of the installation of flow rules that are used for reactive path provisioning. Finally, we measure the response time of the controller when dealing with asynchronous messages.

  @inproceedings{info3-inproceedings-2018-15, abstract = {The increased popularity of Software Defined Networking (SDN) drives the development and research of all involved components. In particular, state-of-the-art SDN controllers are a central part of the resulting network architecture and have a large impact on the performance of the entire network. Hence, it is necessary to quantify various performance indicators of the SDN controller in order to assess its feasibility in the context of different network characteristics such as topology and network size w.r.t. the number of nodes and links. To this end, we extend the open-source controller benchmarking tool OFCProbe with several novel modules that are capable of recording relevant performance metrics and demonstrate their use in the context of benchmarking the ONOS controller. Furthermore, we investigate the impact of different types of topology characteristics on the performance of the controller. In this work, we focus on several performance indicators for SDN controllers. First, the time that the ONOS controller needs to discover the network topology and update its internal representation of the topology. Second, we investigate the duration of the installation of flow rules that are used for reactive path provisioning. Finally, we measure the response time of the controller when dealing with asynchronous messages.}, address = {Hue, Vietnam}, author = {Nguyen-Ngoc, Anh and Raffeck, Simon and Lange, Stanislav and Geissler, Stefan and Zinner, Thomas and Tran-Gia, Phuoc}, booktitle = {7th International Conference on Communications and Electronics (ICCE)}, keywords = {ngn}, month = 7, title = {Benchmarking the ONOS Controller with OFCProbe}, year = 2018 }

  %0 Conference Paper %1 info3-inproceedings-2018-15 %A Nguyen-Ngoc, Anh %A Raffeck, Simon %A Lange, Stanislav %A Geissler, Stefan %A Zinner, Thomas %A Tran-Gia, Phuoc %B 7th International Conference on Communications and Electronics (ICCE) %C Hue, Vietnam %D 2018 %T Benchmarking the ONOS Controller with OFCProbe %X The increased popularity of Software Defined Networking (SDN) drives the development and research of all involved components. In particular, state-of-the-art SDN controllers are a central part of the resulting network architecture and have a large impact on the performance of the entire network. Hence, it is necessary to quantify various performance indicators of the SDN controller in order to assess its feasibility in the context of different network characteristics such as topology and network size w.r.t. the number of nodes and links. To this end, we extend the open-source controller benchmarking tool OFCProbe with several novel modules that are capable of recording relevant performance metrics and demonstrate their use in the context of benchmarking the ONOS controller. Furthermore, we investigate the impact of different types of topology characteristics on the performance of the controller. In this work, we focus on several performance indicators for SDN controllers. First, the time that the ONOS controller needs to discover the network topology and update its internal representation of the topology. Second, we investigate the duration of the installation of flow rules that are used for reactive path provisioning. Finally, we measure the response time of the controller when dealing with asynchronous messages.

• A Discrete-Time Model for Optimizing the Processing Time of Virtualized Network Functions. Zinner, Thomas; Geissler, Stefan; Lange, Stanislav; Gebert, Steffen; Seufert, Michael; Tran-Gia, Phuoc. In Computer Networks, 125, pp. 4–14. 2017.
  • [ Abstract ]
  • [ BibTeX ]
  • [ EndNote ]
  • [ Download ]
  The softwarization of networks promises cost savings and better scalability of network functions by moving functionality from specialized devices into commercial off-the-shelf hardware. Generalized computing hardware offers many degrees of adjustment and tuning, which can affect performance and resource utilization. One of these adjustments are interrupt moderation techniques implemented by modern network interface cards and operating systems. Using these, an administrator can optimize either for low latencies or low CPU overhead for processing of network traffic. In this work, an analytical model that allows the computation of relevant performance metrics like packet processing time and packet loss for generic virtualized network functions running on commodity hardware is presented. Based on this model, impact factors like average packet interarrival time, interarrival time distribution, and duration of the interrupt aggregation interval are studied. Furthermore, we significantly improve the computational tractability of this discrete-time model by proving and leveraging a property regarding its limit behavior. We also demonstrate that using this property does not affect the accuracy of the model in the context of realistic parameter combinations. Finally, the improved runtime for numerical evaluations allows administrators to dynamically adapt their interrupt mitigation settings to changing network conditions by recalculating optimal parameters.

  @article{info3-article-2017-1, abstract = {The softwarization of networks promises cost savings and better scalability of network functions by moving functionality from specialized devices into commercial off-the-shelf hardware. Generalized computing hardware offers many degrees of adjustment and tuning, which can affect performance and resource utilization. One of these adjustments are interrupt moderation techniques implemented by modern network interface cards and operating systems. Using these, an administrator can optimize either for low latencies or low CPU overhead for processing of network traffic. In this work, an analytical model that allows the computation of relevant performance metrics like packet processing time and packet loss for generic virtualized network functions running on commodity hardware is presented. Based on this model, impact factors like average packet interarrival time, interarrival time distribution, and duration of the interrupt aggregation interval are studied. Furthermore, we significantly improve the computational tractability of this discrete-time model by proving and leveraging a property regarding its limit behavior. We also demonstrate that using this property does not affect the accuracy of the model in the context of realistic parameter combinations. Finally, the improved runtime for numerical evaluations allows administrators to dynamically adapt their interrupt mitigation settings to changing network conditions by recalculating optimal parameters.}, author = {Zinner, Thomas and Geissler, Stefan and Lange, Stanislav and Gebert, Steffen and Seufert, Michael and Tran-Gia, Phuoc}, journal = {Computer Networks}, keywords = {sys:hidden:article}, pages = {4-14}, title = {A Discrete-Time Model for Optimizing the Processing Time of Virtualized Network Functions}, volume = 125, year = 2017 }

  %0 Journal Article %1 info3-article-2017-1 %A Zinner, Thomas %A Geissler, Stefan %A Lange, Stanislav %A Gebert, Steffen %A Seufert, Michael %A Tran-Gia, Phuoc %D 2017 %J Computer Networks %P 4-14 %T A Discrete-Time Model for Optimizing the Processing Time of Virtualized Network Functions %V 125 %X The softwarization of networks promises cost savings and better scalability of network functions by moving functionality from specialized devices into commercial off-the-shelf hardware. Generalized computing hardware offers many degrees of adjustment and tuning, which can affect performance and resource utilization. One of these adjustments are interrupt moderation techniques implemented by modern network interface cards and operating systems. Using these, an administrator can optimize either for low latencies or low CPU overhead for processing of network traffic. In this work, an analytical model that allows the computation of relevant performance metrics like packet processing time and packet loss for generic virtualized network functions running on commodity hardware is presented. Based on this model, impact factors like average packet interarrival time, interarrival time distribution, and duration of the interrupt aggregation interval are studied. Furthermore, we significantly improve the computational tractability of this discrete-time model by proving and leveraging a property regarding its limit behavior. We also demonstrate that using this property does not affect the accuracy of the model in the context of realistic parameter combinations. Finally, the improved runtime for numerical evaluations allows administrators to dynamically adapt their interrupt mitigation settings to changing network conditions by recalculating optimal parameters.
• Performance Evaluation of Selective Flow Monitoring in the ONOS Controller. Nguyen-Ngoc, Anh; Lange, Stanislav; Zinner, Thomas; Seufert, Michael; Tran-Gia, Phuoc; Aerts, Nieke; Hock, David. In 4th International Workshop on Management of SDN and NFV Systems (ManSDN/NFV). Tokio, Japan, 2017.
  • [ Abstract ]
  • [ BibTeX ]
  • [ EndNote ]
  One of the benefits when network operators adopt the Software Defined Networking (SDN) paradigm is the ability to monitor the traffic in the network without an additional network management system. Usually, SDN controllers utilize OpenFlow statistics messages in order to regularly gather information about all flows in the network. However, using the same polling interval for all flows does not take into account the heterogeneity of real world traffic and thus results in an imbalance between monitoring accuracy and control plane overhead. In particular, frequent querying results in a high resource consumption at the controller. This work proposes a Selective Flow Monitoring (SFM) mechanism that allows administrators to classify flows according to their individual requirements in terms of monitoring frequency, e.g., less frequent polling of elephant flows and frequent polling of QoS sensitive VoIP connections. We compare the performance of the SFM mechanism with the default monitoring scheme in a testbed featuring the Open Network Operating System (ONOS) controller. In this context, the CPU utilization of the controller is used as performance indicator. After identifying relevant influence factors like the number of flows and switches in the network, we investigate the viability of the approaches in different scenarios. Finally, we provide guidelines regarding their choice.

  @inproceedings{info3-inproceedings-2017-20, abstract = {One of the benefits when network operators adopt the Software Defined Networking (SDN) paradigm is the ability to monitor the traffic in the network without an additional network management system. Usually, SDN controllers utilize OpenFlow statistics messages in order to regularly gather information about all flows in the network. However, using the same polling interval for all flows does not take into account the heterogeneity of real world traffic and thus results in an imbalance between monitoring accuracy and control plane overhead. In particular, frequent querying results in a high resource consumption at the controller. This work proposes a Selective Flow Monitoring (SFM) mechanism that allows administrators to classify flows according to their individual requirements in terms of monitoring frequency, e.g., less frequent polling of elephant flows and frequent polling of QoS sensitive VoIP connections. We compare the performance of the SFM mechanism with the default monitoring scheme in a testbed featuring the Open Network Operating System (ONOS) controller. In this context, the CPU utilization of the controller is used as performance indicator. After identifying relevant influence factors like the number of flows and switches in the network, we investigate the viability of the approaches in different scenarios. Finally, we provide guidelines regarding their choice.}, address = {Tokio, Japan}, author = {Nguyen-Ngoc, Anh and Lange, Stanislav and Zinner, Thomas and Seufert, Michael and Tran-Gia, Phuoc and Aerts, Nieke and Hock, David}, booktitle = {4th International Workshop on Management of SDN and NFV Systems (ManSDN/NFV)}, keywords = {sys:hidden:inproceedings}, month = 11, title = {Performance Evaluation of Selective Flow Monitoring in the ONOS Controller}, year = 2017 }

  %0 Conference Paper %1 info3-inproceedings-2017-20 %A Nguyen-Ngoc, Anh %A Lange, Stanislav %A Zinner, Thomas %A Seufert, Michael %A Tran-Gia, Phuoc %A Aerts, Nieke %A Hock, David %B 4th International Workshop on Management of SDN and NFV Systems (ManSDN/NFV) %C Tokio, Japan %D 2017 %T Performance Evaluation of Selective Flow Monitoring in the ONOS Controller %X One of the benefits when network operators adopt the Software Defined Networking (SDN) paradigm is the ability to monitor the traffic in the network without an additional network management system. Usually, SDN controllers utilize OpenFlow statistics messages in order to regularly gather information about all flows in the network. However, using the same polling interval for all flows does not take into account the heterogeneity of real world traffic and thus results in an imbalance between monitoring accuracy and control plane overhead. In particular, frequent querying results in a high resource consumption at the controller. This work proposes a Selective Flow Monitoring (SFM) mechanism that allows administrators to classify flows according to their individual requirements in terms of monitoring frequency, e.g., less frequent polling of elephant flows and frequent polling of QoS sensitive VoIP connections. We compare the performance of the SFM mechanism with the default monitoring scheme in a testbed featuring the Open Network Operating System (ONOS) controller. In this context, the CPU utilization of the controller is used as performance indicator. After identifying relevant influence factors like the number of flows and switches in the network, we investigate the viability of the approaches in different scenarios. Finally, we provide guidelines regarding their choice.
• A Multi-Objective Heuristic for the Optimization of Virtual Network Function Chain Placement. Lange, Stanislav; Grigorjew, Alexej; Zinner, Thomas; Tran-Gia, Phuoc; Jarschel, Michael. In 29th International Teletraffic Congress (ITC). 2017.
  • [ BibTeX ]
  • [ EndNote ]
  • [ Download ]
  @inproceedings{info3-inproceedings-2017-16, author = {Lange, Stanislav and Grigorjew, Alexej and Zinner, Thomas and Tran-Gia, Phuoc and Jarschel, Michael}, booktitle = {29th International Teletraffic Congress (ITC)}, keywords = {sendate}, month = 9, title = {A Multi-Objective Heuristic for the Optimization of Virtual Network Function Chain Placement}, year = 2017 }

  %0 Conference Paper %1 info3-inproceedings-2017-16 %A Lange, Stanislav %A Grigorjew, Alexej %A Zinner, Thomas %A Tran-Gia, Phuoc %A Jarschel, Michael %B 29th International Teletraffic Congress (ITC) %D 2017 %T A Multi-Objective Heuristic for the Optimization of Virtual Network Function Chain Placement
• Automated Decision Making based on Pareto Frontiers in the Context of Service Placement in Networks. Seufert, Michael; Lange, Stanislav; Meixner, Markus. In 29th International Teletraffic Congress (ITC). Genoa, Italy, 2017.
  • [ Abstract ]
  • [ BibTeX ]
  • [ EndNote ]
  • [ Download ]
  Virtualization paradigms like cloud computing, software defined networking (SDN), and network functions virtualization (NFV) provide advantages with respect to aspects like flexibility, costs, and scalability. However, management and orchestration of the resulting networks also introduce new challenges. The placement of services, such as virtual machines~(VMs), virtualized network functions~(VNFs), or SDN controllers, is a multi-objective optimization task that confronts operators with a multitude of possible solutions that are incomparable among each other. The goal of this work is to investigate mechanisms that enable automated decision making between such multi-dimensional solutions. To this end, we investigate techniques from the domain of multi-attribute decision making that aggregate the performance of placements to a single numeric score. A comparison between resulting rankings of placements shows that many techniques produce similar results. Hence, placements that achieve good rankings according to many approaches might be viable candidates in the context of automated decision making. In order to illustrate the functionality of the different scoring mechanisms, we perform a case study on a single network graph and a fixed number of objectives and service instances. Additionally, we present aggregated results from broad evaluations on the Internet Topology Zoo and a larger number of objectives as well as varying numbers of service instances. These allow making more reliable statements about the mechanisms' performance and agreement.

  @inproceedings{info3-article-2017-6, abstract = {Virtualization paradigms like cloud computing, software defined networking (SDN), and network functions virtualization (NFV) provide advantages with respect to aspects like flexibility, costs, and scalability. However, management and orchestration of the resulting networks also introduce new challenges. The placement of services, such as virtual machines (VMs), virtualized network functions (VNFs), or SDN controllers, is a multi-objective optimization task that confronts operators with a multitude of possible solutions that are incomparable among each other. The goal of this work is to investigate mechanisms that enable automated decision making between such multi-dimensional solutions. To this end, we investigate techniques from the domain of multi-attribute decision making that aggregate the performance of placements to a single numeric score. A comparison between resulting rankings of placements shows that many techniques produce similar results. Hence, placements that achieve good rankings according to many approaches might be viable candidates in the context of automated decision making. In order to illustrate the functionality of the different scoring mechanisms, we perform a case study on a single network graph and a fixed number of objectives and service instances. Additionally, we present aggregated results from broad evaluations on the Internet Topology Zoo and a larger number of objectives as well as varying numbers of service instances. These allow making more reliable statements about the mechanisms' performance and agreement.}, address = {Genoa, Italy}, author = {Seufert, Michael and Lange, Stanislav and Meixner, Markus}, booktitle = {29th International Teletraffic Congress (ITC)}, keywords = {can}, month = 9, title = {Automated Decision Making based on Pareto Frontiers in the Context of Service Placement in Networks}, year = 2017 }

  %0 Conference Paper %1 info3-article-2017-6 %A Seufert, Michael %A Lange, Stanislav %A Meixner, Markus %B 29th International Teletraffic Congress (ITC) %C Genoa, Italy %D 2017 %T Automated Decision Making based on Pareto Frontiers in the Context of Service Placement in Networks %X Virtualization paradigms like cloud computing, software defined networking (SDN), and network functions virtualization (NFV) provide advantages with respect to aspects like flexibility, costs, and scalability. However, management and orchestration of the resulting networks also introduce new challenges. The placement of services, such as virtual machines~(VMs), virtualized network functions~(VNFs), or SDN controllers, is a multi-objective optimization task that confronts operators with a multitude of possible solutions that are incomparable among each other. The goal of this work is to investigate mechanisms that enable automated decision making between such multi-dimensional solutions. To this end, we investigate techniques from the domain of multi-attribute decision making that aggregate the performance of placements to a single numeric score. A comparison between resulting rankings of placements shows that many techniques produce similar results. Hence, placements that achieve good rankings according to many approaches might be viable candidates in the context of automated decision making. In order to illustrate the functionality of the different scoring mechanisms, we perform a case study on a single network graph and a fixed number of objectives and service instances. Additionally, we present aggregated results from broad evaluations on the Internet Topology Zoo and a larger number of objectives as well as varying numbers of service instances. These allow making more reliable statements about the mechanisms' performance and agreement.

• Performance Evaluation Mechanisms for FlowMod Message Processing in OpenFlow Switches. Nguyen-Ngoc, Anh; Lange, Stanislav; Gebert, Steffen; Zinner, Thomas; Tran-Gia, Phuoc; Jarschel, Michael. In IEEE Sixth International Conference on Communications and Electronics. Ha Long City, Vietnam, 2016.
  • [ Abstract ]
  • [ BibTeX ]
  • [ EndNote ]
  • [ Download ]
  Network operators can benefit in terms of flexibility, cost, and vendor-independence when adopting the Software Defined Networking (SDN) paradigm. In many scenarios, the SDN controller orders the installation of new flow table entries in the switches it manages. Since such operations are handled in the slow path of the switches, the corresponding processing times constitute an important performance indicator for switches. This work focuses on a comparison of two mechanisms for evaluating the performance of OpenFlow switches with respect to the processing time of FlowMod messages. These mechanisms are characterized by different degrees of accuracy, cost, complexity, and the capability of performing measurements at run time. The first mechanism is based on the Spirent C1 dedicated testing platform, while the other uses a software module for the OpenDaylight controller. We assess their capabilities with respect to the abovementioned characteristics and quantify their accuracy by means of wiretaps that provide a ground truth regarding the measured processing times. By using three different switches in the experiments, it is possible to distinguish between hardware specific side-effects and general phenomena.

  @inproceedings{info3-inproceedings-2016-7, abstract = {Network operators can benefit in terms of flexibility, cost, and vendor-independence when adopting the Software Defined Networking (SDN) paradigm. In many scenarios, the SDN controller orders the installation of new flow table entries in the switches it manages. Since such operations are handled in the slow path of the switches, the corresponding processing times constitute an important performance indicator for switches. This work focuses on a comparison of two mechanisms for evaluating the performance of OpenFlow switches with respect to the processing time of FlowMod messages. These mechanisms are characterized by different degrees of accuracy, cost, complexity, and the capability of performing measurements at run time. The first mechanism is based on the Spirent C1 dedicated testing platform, while the other uses a software module for the OpenDaylight controller. We assess their capabilities with respect to the abovementioned characteristics and quantify their accuracy by means of wiretaps that provide a ground truth regarding the measured processing times. By using three different switches in the experiments, it is possible to distinguish between hardware specific side-effects and general phenomena.}, address = {Ha Long City, Vietnam}, author = {Nguyen-Ngoc, Anh and Lange, Stanislav and Gebert, Steffen and Zinner, Thomas and Tran-Gia, Phuoc and Jarschel, Michael}, booktitle = {IEEE Sixth International Conference on Communications and Electronics}, keywords = {sendate}, month = 7, title = {Performance Evaluation Mechanisms for FlowMod Message Processing in OpenFlow Switches}, year = 2016 }

  %0 Conference Paper %1 info3-inproceedings-2016-7 %A Nguyen-Ngoc, Anh %A Lange, Stanislav %A Gebert, Steffen %A Zinner, Thomas %A Tran-Gia, Phuoc %A Jarschel, Michael %B IEEE Sixth International Conference on Communications and Electronics %C Ha Long City, Vietnam %D 2016 %T Performance Evaluation Mechanisms for FlowMod Message Processing in OpenFlow Switches %X Network operators can benefit in terms of flexibility, cost, and vendor-independence when adopting the Software Defined Networking (SDN) paradigm. In many scenarios, the SDN controller orders the installation of new flow table entries in the switches it manages. Since such operations are handled in the slow path of the switches, the corresponding processing times constitute an important performance indicator for switches. This work focuses on a comparison of two mechanisms for evaluating the performance of OpenFlow switches with respect to the processing time of FlowMod messages. These mechanisms are characterized by different degrees of accuracy, cost, complexity, and the capability of performing measurements at run time. The first mechanism is based on the Spirent C1 dedicated testing platform, while the other uses a software module for the OpenDaylight controller. We assess their capabilities with respect to the abovementioned characteristics and quantify their accuracy by means of wiretaps that provide a ground truth regarding the measured processing times. By using three different switches in the experiments, it is possible to distinguish between hardware specific side-effects and general phenomena.
• More than Topology: Joint Topology and Attribute Sampling and Generation of Social Network Graphs. Seufert, Michael; Lange, Stanislav; Hoßfeld, Tobias. In Computer Communications, 73(Part B), pp. 176–187. 2016.
  • [ Abstract ]
  • [ BibTeX ]
  • [ EndNote ]
  • [ Download ]
  Graph sampling refers to the process of deriving a small subset of nodes from a possibly huge graph in order to estimate properties of the whole graph from examining the sample. Whereas topological properties can already be obtained accurately by sampling, current approaches do not take possibly hidden dependencies between node topology and attributes into account. Especially in the context of online social networks, node attributes are of importance as they correspond to properties of the social network's users. Therefore, existing sampling algorithms can be extended to attribute sampling, but still lack the capturing of structural properties. Analyzing topology (e.g., node degree, clustering coefficient) and attribute properties (e.g., age, location) jointly can provide valuable insights into the social network and allows for a better understanding of social processes. As major contribution, this work proposes a novel sampling algorithm which provides unbiased and reliable estimates of joint topological and attribute based graph properties in a resource efficient fashion. Furthermore, the obtained samples allow for the generation of synthetic graphs, which show high similarity to the original graph with respect to topology and attributes. The proposed sampling and generation algorithms are evaluated on real world social network graphs, for which they demonstrate to be effective.

  @article{info3-article-2016-1, abstract = {Graph sampling refers to the process of deriving a small subset of nodes from a possibly huge graph in order to estimate properties of the whole graph from examining the sample. Whereas topological properties can already be obtained accurately by sampling, current approaches do not take possibly hidden dependencies between node topology and attributes into account. Especially in the context of online social networks, node attributes are of importance as they correspond to properties of the social network's users. Therefore, existing sampling algorithms can be extended to attribute sampling, but still lack the capturing of structural properties. Analyzing topology (e.g., node degree, clustering coefficient) and attribute properties (e.g., age, location) jointly can provide valuable insights into the social network and allows for a better understanding of social processes. As major contribution, this work proposes a novel sampling algorithm which provides unbiased and reliable estimates of joint topological and attribute based graph properties in a resource efficient fashion. Furthermore, the obtained samples allow for the generation of synthetic graphs, which show high similarity to the original graph with respect to topology and attributes. The proposed sampling and generation algorithms are evaluated on real world social network graphs, for which they demonstrate to be effective.}, author = {Seufert, Michael and Lange, Stanislav and Hoßfeld, Tobias}, journal = {Computer Communications}, keywords = {can}, month = 1, number = {Part B}, pages = {176-187}, title = {More than Topology: Joint Topology and Attribute Sampling and Generation of Social Network Graphs}, volume = 73, year = 2016 }

  %0 Journal Article %1 info3-article-2016-1 %A Seufert, Michael %A Lange, Stanislav %A Hoßfeld, Tobias %D 2016 %J Computer Communications %N Part B %P 176-187 %T More than Topology: Joint Topology and Attribute Sampling and Generation of Social Network Graphs %V 73 %X Graph sampling refers to the process of deriving a small subset of nodes from a possibly huge graph in order to estimate properties of the whole graph from examining the sample. Whereas topological properties can already be obtained accurately by sampling, current approaches do not take possibly hidden dependencies between node topology and attributes into account. Especially in the context of online social networks, node attributes are of importance as they correspond to properties of the social network's users. Therefore, existing sampling algorithms can be extended to attribute sampling, but still lack the capturing of structural properties. Analyzing topology (e.g., node degree, clustering coefficient) and attribute properties (e.g., age, location) jointly can provide valuable insights into the social network and allows for a better understanding of social processes. As major contribution, this work proposes a novel sampling algorithm which provides unbiased and reliable estimates of joint topological and attribute based graph properties in a resource efficient fashion. Furthermore, the obtained samples allow for the generation of synthetic graphs, which show high similarity to the original graph with respect to topology and attributes. The proposed sampling and generation algorithms are evaluated on real world social network graphs, for which they demonstrate to be effective.
• Demonstrating a Personalized Secure-By-Default Bring Your Own Device Solution Based on Software Defined Networking. Gebert, Steffen; Zinner, Thomas; Gray, Nicholas; Durner, Raphael; Lorenz, Class; Lange, Stanislav. Würzburg, Germany, 2016, September.
  • [ Abstract ]
  • [ BibTeX ]
  • [ EndNote ]
  • [ Download ]
  Network virtualization is one classical use-case for Software Defined Networks (SDN). By programmatically instantiating virtual networks, traffic from one or more devices can be separated or connectivity can be established as needed. S- BYOD, which is presented in this demonstration, applies the SDN concept to Bring Your Own Device (BYOD) scenarios and offers personalized virtual networks that are set up and extended on demand. This is done once the user authenticates, activates access to additional applications, or as soon as applications scale out and involve more servers. The described proof-of-concept implementation explores, to what degree an agent-less BYOD solution, based only on SDN, can lower the attack surface by explicit user opt-ins for particular services. Further, an assessment of the number of required rules within the flow tables of switches completes this work.

  @misc{info3-demo-2016-1, abstract = {Network virtualization is one classical use-case for Software Defined Networks (SDN). By programmatically instantiating virtual networks, traffic from one or more devices can be separated or connectivity can be established as needed. S- BYOD, which is presented in this demonstration, applies the SDN concept to Bring Your Own Device (BYOD) scenarios and offers personalized virtual networks that are set up and extended on demand. This is done once the user authenticates, activates access to additional applications, or as soon as applications scale out and involve more servers. The described proof-of-concept implementation explores, to what degree an agent-less BYOD solution, based only on SDN, can lower the attack surface by explicit user opt-ins for particular services. Further, an assessment of the number of required rules within the flow tables of switches completes this work.}, address = {Würzburg, Germany}, author = {Gebert, Steffen and Zinner, Thomas and Gray, Nicholas and Durner, Raphael and Lorenz, Class and Lange, Stanislav}, booktitle = {28th International Teletraffic Congress (ITC 28)}, keywords = {sardine}, month = 9, title = {Demonstrating a Personalized Secure-By-Default Bring Your Own Device Solution Based on Software Defined Networking}, year = 2016 }

  %0 Generic %1 info3-demo-2016-1 %A Gebert, Steffen %A Zinner, Thomas %A Gray, Nicholas %A Durner, Raphael %A Lorenz, Class %A Lange, Stanislav %B 28th International Teletraffic Congress (ITC 28) %C Würzburg, Germany %D 2016 %T Demonstrating a Personalized Secure-By-Default Bring Your Own Device Solution Based on Software Defined Networking %X Network virtualization is one classical use-case for Software Defined Networks (SDN). By programmatically instantiating virtual networks, traffic from one or more devices can be separated or connectivity can be established as needed. S- BYOD, which is presented in this demonstration, applies the SDN concept to Bring Your Own Device (BYOD) scenarios and offers personalized virtual networks that are set up and extended on demand. This is done once the user authenticates, activates access to additional applications, or as soon as applications scale out and involve more servers. The described proof-of-concept implementation explores, to what degree an agent-less BYOD solution, based only on SDN, can lower the attack surface by explicit user opt-ins for particular services. Further, an assessment of the number of required rules within the flow tables of switches completes this work.
• Performance Modeling of Softwarized Network Functions Using Discrete-Time Analysis. Gebert, Steffen; Zinner, Thomas; Lange, Stanislav; Schwartz, Christian; Tran-Gia, Phuoc. In 28th International Teletraffic Congress (ITC). Würzburg, Germany, 2016.
  • [ Abstract ]
  • [ BibTeX ]
  • [ EndNote ]
  • [ Download ]
  The softwarization of networks promises cost savings and better scalability of network functions by moving functionality from specialized devices into commercial off-the-shelf hardware. Generalized computing hardware offers many degrees for adjustment and tuning, which can affect performance and resource utilization. One of these adjustments are the interrupt mitigation techniques implemented by modern network interface cards and operating systems. Using these, an administrator can optimize either lower latencies or lower CPU overhead for processing of network traffic. In this work, an analytical model that allows computing relevant performance metrics like the packet processing time and the packet loss for generic virtualized network functions running on commodity hardware is developed. The applicability of the model is shown by comparing its outcome with measurements conducted in a local testbed. Based on this model, impact factors like the average packet interarrival time, the interarrival time distribution, and the duration of the interrupt aggregation interval are studied.

  @inproceedings{info3-inproceedings-2016-4, abstract = {The softwarization of networks promises cost savings and better scalability of network functions by moving functionality from specialized devices into commercial off-the-shelf hardware. Generalized computing hardware offers many degrees for adjustment and tuning, which can affect performance and resource utilization. One of these adjustments are the interrupt mitigation techniques implemented by modern network interface cards and operating systems. Using these, an administrator can optimize either lower latencies or lower CPU overhead for processing of network traffic. In this work, an analytical model that allows computing relevant performance metrics like the packet processing time and the packet loss for generic virtualized network functions running on commodity hardware is developed. The applicability of the model is shown by comparing its outcome with measurements conducted in a local testbed. Based on this model, impact factors like the average packet interarrival time, the interarrival time distribution, and the duration of the interrupt aggregation interval are studied.}, address = {Würzburg, Germany}, author = {Gebert, Steffen and Zinner, Thomas and Lange, Stanislav and Schwartz, Christian and Tran-Gia, Phuoc}, booktitle = {28th International Teletraffic Congress (ITC)}, keywords = {sendate}, month = 9, title = {Performance Modeling of Softwarized Network Functions Using Discrete-Time Analysis}, year = 2016 }

  %0 Conference Paper %1 info3-inproceedings-2016-4 %A Gebert, Steffen %A Zinner, Thomas %A Lange, Stanislav %A Schwartz, Christian %A Tran-Gia, Phuoc %B 28th International Teletraffic Congress (ITC) %C Würzburg, Germany %D 2016 %T Performance Modeling of Softwarized Network Functions Using Discrete-Time Analysis %X The softwarization of networks promises cost savings and better scalability of network functions by moving functionality from specialized devices into commercial off-the-shelf hardware. Generalized computing hardware offers many degrees for adjustment and tuning, which can affect performance and resource utilization. One of these adjustments are the interrupt mitigation techniques implemented by modern network interface cards and operating systems. Using these, an administrator can optimize either lower latencies or lower CPU overhead for processing of network traffic. In this work, an analytical model that allows computing relevant performance metrics like the packet processing time and the packet loss for generic virtualized network functions running on commodity hardware is developed. The applicability of the model is shown by comparing its outcome with measurements conducted in a local testbed. Based on this model, impact factors like the average packet interarrival time, the interarrival time distribution, and the duration of the interrupt aggregation interval are studied.
• ZOOM: Lightweight SDN-based Elephant Detection. Gebert, Steffen; Geissler, Stefan; Zinner, Thomas; Nguyen-Ngoc, Anh; Lange, Stanislav; Tran-Gia, Phuoc. In First International Workshop on Programmability for Cloud Networks and Applications (PROCON). Würzburg, Germany, 2016.
  • [ Abstract ]
  • [ BibTeX ]
  • [ EndNote ]
  • [ Download ]
  Software Defined Networking (SDN) offers a holistic view of the network through a centralized control plane. Consequently, routing decisions can be made based on global knowledge about the network topology as well as its current state. As long living flows are suitable candidates for rerouting, their detection is crucial for efficient flow based traffic management. This work proposes the ZOOM algorithm for elephant detection in SDN networks. To this end, ZOOM follows a very lightweight approach that only uses packet counters implemented by OpenFlow switches and thus does not require any additional hardware. By exploiting this feature of OpenFlow switches, ZOOM allows lightweight and cost-effective elephant detection.

  @inproceedings{info3-inproceedings-2016-8, abstract = {Software Defined Networking (SDN) offers a holistic view of the network through a centralized control plane. Consequently, routing decisions can be made based on global knowledge about the network topology as well as its current state. As long living flows are suitable candidates for rerouting, their detection is crucial for efficient flow based traffic management. This work proposes the ZOOM algorithm for elephant detection in SDN networks. To this end, ZOOM follows a very lightweight approach that only uses packet counters implemented by OpenFlow switches and thus does not require any additional hardware. By exploiting this feature of OpenFlow switches, ZOOM allows lightweight and cost-effective elephant detection.}, address = {Würzburg, Germany}, author = {Gebert, Steffen and Geissler, Stefan and Zinner, Thomas and Nguyen-Ngoc, Anh and Lange, Stanislav and Tran-Gia, Phuoc}, booktitle = {First International Workshop on Programmability for Cloud Networks and Applications (PROCON)}, keywords = {sardine}, month = 9, title = {ZOOM: Lightweight SDN-based Elephant Detection}, year = 2016 }

  %0 Conference Paper %1 info3-inproceedings-2016-8 %A Gebert, Steffen %A Geissler, Stefan %A Zinner, Thomas %A Nguyen-Ngoc, Anh %A Lange, Stanislav %A Tran-Gia, Phuoc %B First International Workshop on Programmability for Cloud Networks and Applications (PROCON) %C Würzburg, Germany %D 2016 %T ZOOM: Lightweight SDN-based Elephant Detection %X Software Defined Networking (SDN) offers a holistic view of the network through a centralized control plane. Consequently, routing decisions can be made based on global knowledge about the network topology as well as its current state. As long living flows are suitable candidates for rerouting, their detection is crucial for efficient flow based traffic management. This work proposes the ZOOM algorithm for elephant detection in SDN networks. To this end, ZOOM follows a very lightweight approach that only uses packet counters implemented by OpenFlow switches and thus does not require any additional hardware. By exploiting this feature of OpenFlow switches, ZOOM allows lightweight and cost-effective elephant detection.
• Automated Decision Making Methods for the Multi-objective Optimization Task of Cloud Service Placement. Seufert, Michael; Lange, Stanislav; Meixner, Markus. In 1st International Workshop on Programmability for Cloud Networks and Applications (PROCON). Würzburg, Germany, 2016.
  • [ Abstract ]
  • [ BibTeX ]
  • [ EndNote ]
  • [ Download ]
  The network functions virtualization (NFV) paradigm provides advantages with respect to aspects like flexibility, costs, and scalability of networks. However, management and orchestration of the resulting networks also introduce new challenges. The placement of services and virtualized network functions (VNFs) is a multi-objective optimization task that confronts operators with a multitude of possible solutions that are incomparable among each other. The goal of this work is to investigate mechanisms that enable automated decision making between such multi dimensional solutions. To this end, we investigate techniques from the domain of multi attribute decision making that aggregate the performance of placements to a single numeric score. A comparison between resulting rankings of placements shows that many techniques produce similar results. Hence, placements that achieve good rankings according to many approaches might be viable candidates in the context of automated decision making.

  @inproceedings{info3-inproceedings-2016-12, abstract = {The network functions virtualization (NFV) paradigm provides advantages with respect to aspects like flexibility, costs, and scalability of networks. However, management and orchestration of the resulting networks also introduce new challenges. The placement of services and virtualized network functions (VNFs) is a multi-objective optimization task that confronts operators with a multitude of possible solutions that are incomparable among each other. The goal of this work is to investigate mechanisms that enable automated decision making between such multi dimensional solutions. To this end, we investigate techniques from the domain of multi attribute decision making that aggregate the performance of placements to a single numeric score. A comparison between resulting rankings of placements shows that many techniques produce similar results. Hence, placements that achieve good rankings according to many approaches might be viable candidates in the context of automated decision making.}, address = {Würzburg, Germany}, author = {Seufert, Michael and Lange, Stanislav and Meixner, Markus}, booktitle = {1st International Workshop on Programmability for Cloud Networks and Applications (PROCON)}, keywords = {can}, month = 9, title = {Automated Decision Making Methods for the Multi-objective Optimization Task of Cloud Service Placement}, year = 2016 }

  %0 Conference Paper %1 info3-inproceedings-2016-12 %A Seufert, Michael %A Lange, Stanislav %A Meixner, Markus %B 1st International Workshop on Programmability for Cloud Networks and Applications (PROCON) %C Würzburg, Germany %D 2016 %T Automated Decision Making Methods for the Multi-objective Optimization Task of Cloud Service Placement %X The network functions virtualization (NFV) paradigm provides advantages with respect to aspects like flexibility, costs, and scalability of networks. However, management and orchestration of the resulting networks also introduce new challenges. The placement of services and virtualized network functions (VNFs) is a multi-objective optimization task that confronts operators with a multitude of possible solutions that are incomparable among each other. The goal of this work is to investigate mechanisms that enable automated decision making between such multi dimensional solutions. To this end, we investigate techniques from the domain of multi attribute decision making that aggregate the performance of placements to a single numeric score. A comparison between resulting rankings of placements shows that many techniques produce similar results. Hence, placements that achieve good rankings according to many approaches might be viable candidates in the context of automated decision making.
• Discrete-Time Analysis: Deriving the Distribution of the Number of Events in an Arbitrarily Distributed Interval. Technical Report (498), . Gebert, Steffen; Zinner, Thomas; Lange, Stanislav; Schwartz, Christian; Tran-Gia, Phuoc. University of Wuerzburg, 2016.
  • [ BibTeX ]
  • [ EndNote ]
  • [ Download ]
  @techreport{info3-techreport-2016-1, author = {Gebert, Steffen and Zinner, Thomas and Lange, Stanislav and Schwartz, Christian and Tran-Gia, Phuoc}, institution = {University of Wuerzburg}, keywords = {sys:hidden:techreport}, month = 6, number = 498, title = {Discrete-Time Analysis: Deriving the Distribution of the Number of Events in an Arbitrarily Distributed Interval}, year = 2016 }

  %0 Report %1 info3-techreport-2016-1 %A Gebert, Steffen %A Zinner, Thomas %A Lange, Stanislav %A Schwartz, Christian %A Tran-Gia, Phuoc %D 2016 %N 498 %T Discrete-Time Analysis: Deriving the Distribution of the Number of Events in an Arbitrarily Distributed Interval
• Processing Time Comparison of a Hardware-Based Firewall and its Virtualized Counterpart. Gebert, Steffen; Müssig, Alexander; Lange, Stanislav; Zinner, Thomas; Gray, Nicholas; Tran-Gia, Phuoc. In 8th EAI International Conference on Mobile Networks and Management (MONAMI 2016). Abu Dhabi, United Arab Emirates, 2016.
  • [ BibTeX ]
  • [ EndNote ]
  • [ Download ]
  @inproceedings{info3-inproceedings-2016-19, address = {Abu Dhabi, United Arab Emirates}, author = {Gebert, Steffen and Müssig, Alexander and Lange, Stanislav and Zinner, Thomas and Gray, Nicholas and Tran-Gia, Phuoc}, booktitle = {8th EAI International Conference on Mobile Networks and Management (MONAMI 2016)}, keywords = {sardine}, month = 10, title = {Processing Time Comparison of a Hardware-Based Firewall and its Virtualized Counterpart}, year = 2016 }

  %0 Conference Paper %1 info3-inproceedings-2016-19 %A Gebert, Steffen %A Müssig, Alexander %A Lange, Stanislav %A Zinner, Thomas %A Gray, Nicholas %A Tran-Gia, Phuoc %B 8th EAI International Conference on Mobile Networks and Management (MONAMI 2016) %C Abu Dhabi, United Arab Emirates %D 2016 %T Processing Time Comparison of a Hardware-Based Firewall and its Virtualized Counterpart

• Performance Benchmarking of a Software-Based LTE SGW. Lange, Stanislav; Nguyen-Ngoc, Anh; Gebert, Steffen; Zinner, Thomas; Jarschel, Michael; Koepsel, Andreas; Sune, Marc; Raumer, Daniel; Gallenmüller, Sebastian; Carle, Georg; Tran-Gia, Phuoc. In 2nd International Workshop on Management of SDN and NFV Systems. Barcelona, Spain, 2015.
  • [ Abstract ]
  • [ BibTeX ]
  • [ EndNote ]
  Network Functions Virtualization (NFV) is a concept that aims at providing network operators with benefits in terms of cost, flexibility, and vendor independence by utilizing virtualization techniques to run network functions as software on commercial off-the-shelf (COTS) hardware. In contrast, prior solutions rely on specialized hardware for each function. Performance evaluation of such systems usually requires a dedicated testbed for each individual component. Rather than analyzing these proprietary black-box components, Virtualized Network Functions (VNFs) are pieces of software that run on COTS hardware and whose properties can be investigated in a generic testbed. However, depending on the underlying hardware, operating system, and implementation, VNFs might behave differently. Therefore, mechanisms for the performance evaluation of VNFs should be similar to benchmarking of software, where different implementations are compared by applying them to predefined test cases and scenarios. This work presents a first step towards a benchmarking framework for VNFs. Given two different implementations of a VNF that acts as LTE Serving Gateway (SGW), influence factors and key performance indicators are identified and a comparison between the two mechanisms is drawn.

  @inproceedings{info3-inproceedings-2015-533, abstract = {Network Functions Virtualization (NFV) is a concept that aims at providing network operators with benefits in terms of cost, flexibility, and vendor independence by utilizing virtualization techniques to run network functions as software on commercial off-the-shelf (COTS) hardware. In contrast, prior solutions rely on specialized hardware for each function. Performance evaluation of such systems usually requires a dedicated testbed for each individual component. Rather than analyzing these proprietary black-box components, Virtualized Network Functions (VNFs) are pieces of software that run on COTS hardware and whose properties can be investigated in a generic testbed. However, depending on the underlying hardware, operating system, and implementation, VNFs might behave differently. Therefore, mechanisms for the performance evaluation of VNFs should be similar to benchmarking of software, where different implementations are compared by applying them to predefined test cases and scenarios. This work presents a first step towards a benchmarking framework for VNFs. Given two different implementations of a VNF that acts as LTE Serving Gateway (SGW), influence factors and key performance indicators are identified and a comparison between the two mechanisms is drawn.}, address = {Barcelona, Spain}, author = {Lange, Stanislav and Nguyen-Ngoc, Anh and Gebert, Steffen and Zinner, Thomas and Jarschel, Michael and Koepsel, Andreas and Sune, Marc and Raumer, Daniel and Gallenmüller, Sebastian and Carle, Georg and Tran-Gia, Phuoc}, booktitle = {2nd International Workshop on Management of SDN and NFV Systems}, keywords = {sys:hidden:inproceedings}, month = 11, title = {Performance Benchmarking of a Software-Based LTE SGW}, year = 2015 }

  %0 Conference Paper %1 info3-inproceedings-2015-533 %A Lange, Stanislav %A Nguyen-Ngoc, Anh %A Gebert, Steffen %A Zinner, Thomas %A Jarschel, Michael %A Koepsel, Andreas %A Sune, Marc %A Raumer, Daniel %A Gallenmüller, Sebastian %A Carle, Georg %A Tran-Gia, Phuoc %B 2nd International Workshop on Management of SDN and NFV Systems %C Barcelona, Spain %D 2015 %T Performance Benchmarking of a Software-Based LTE SGW %X Network Functions Virtualization (NFV) is a concept that aims at providing network operators with benefits in terms of cost, flexibility, and vendor independence by utilizing virtualization techniques to run network functions as software on commercial off-the-shelf (COTS) hardware. In contrast, prior solutions rely on specialized hardware for each function. Performance evaluation of such systems usually requires a dedicated testbed for each individual component. Rather than analyzing these proprietary black-box components, Virtualized Network Functions (VNFs) are pieces of software that run on COTS hardware and whose properties can be investigated in a generic testbed. However, depending on the underlying hardware, operating system, and implementation, VNFs might behave differently. Therefore, mechanisms for the performance evaluation of VNFs should be similar to benchmarking of software, where different implementations are compared by applying them to predefined test cases and scenarios. This work presents a first step towards a benchmarking framework for VNFs. Given two different implementations of a VNF that acts as LTE Serving Gateway (SGW), influence factors and key performance indicators are identified and a comparison between the two mechanisms is drawn.
• Specialized Heuristics for the Controller Placement Problem in Large Scale SDN Networks. Lange, Stanislav; Gebert, Steffen; Spoerhase, Joachim; Rygielski, Piotr; Zinner, Thomas; Kounev, Samuel; Tran-Gia, Phuoc. In International Teletraffic Congress (ITC 27). Ghent, Belgium, 2015.
  • [ Abstract ]
  • [ BibTeX ]
  • [ EndNote ]
  • [ Download ]
  The Software Defined Networking~(SDN) concept introduces a paradigm shift in the networking world towards an externalized control plane which is logically centralized. When designing an SDN-based WAN architecture, it is of vital importance to find a feasible solution to the controller placement problem, i.e., to decide where to position a limited amount of resources within the network. In addition to time-independent constraints regarding aspects like scalability, resilience, and control plane communication delays, dynamically changing network conditions like traffic patterns or bandwidth demands need to be considered as well. Consequently, such dynamic environments call for a regular and fast recalculation of placements in order to adapt to the current situation in a timely manner. While an exhaustive evaluation of all possible solutions can be performed within a practically feasible time frame for small and medium-sized networks, such an approach is out of scope for large problem instances which have significantly higher time and memory requirements. Therefore, this work investigates a specialized heuristic, which takes into account a particular set of optimization objectives and returns solutions representing the possible trade-offs between them. Due to its low computation time and acceptable margin of error, this heuristic can be employed by automatic decision systems operating in dynamic environments.

  @inproceedings{info3-inproceedings-2015-519, abstract = {The Software Defined Networking (SDN) concept introduces a paradigm shift in the networking world towards an externalized control plane which is logically centralized. When designing an SDN-based WAN architecture, it is of vital importance to find a feasible solution to the controller placement problem, i.e., to decide where to position a limited amount of resources within the network. In addition to time-independent constraints regarding aspects like scalability, resilience, and control plane communication delays, dynamically changing network conditions like traffic patterns or bandwidth demands need to be considered as well. Consequently, such dynamic environments call for a regular and fast recalculation of placements in order to adapt to the current situation in a timely manner. While an exhaustive evaluation of all possible solutions can be performed within a practically feasible time frame for small and medium-sized networks, such an approach is out of scope for large problem instances which have significantly higher time and memory requirements. Therefore, this work investigates a specialized heuristic, which takes into account a particular set of optimization objectives and returns solutions representing the possible trade-offs between them. Due to its low computation time and acceptable margin of error, this heuristic can be employed by automatic decision systems operating in dynamic environments.}, address = {Ghent, Belgium}, author = {Lange, Stanislav and Gebert, Steffen and Spoerhase, Joachim and Rygielski, Piotr and Zinner, Thomas and Kounev, Samuel and Tran-Gia, Phuoc}, booktitle = {International Teletraffic Congress (ITC 27)}, keywords = {poco}, month = 9, title = {Specialized Heuristics for the Controller Placement Problem in Large Scale SDN Networks}, year = 2015 }

  %0 Conference Paper %1 info3-inproceedings-2015-519 %A Lange, Stanislav %A Gebert, Steffen %A Spoerhase, Joachim %A Rygielski, Piotr %A Zinner, Thomas %A Kounev, Samuel %A Tran-Gia, Phuoc %B International Teletraffic Congress (ITC 27) %C Ghent, Belgium %D 2015 %T Specialized Heuristics for the Controller Placement Problem in Large Scale SDN Networks %X The Software Defined Networking~(SDN) concept introduces a paradigm shift in the networking world towards an externalized control plane which is logically centralized. When designing an SDN-based WAN architecture, it is of vital importance to find a feasible solution to the controller placement problem, i.e., to decide where to position a limited amount of resources within the network. In addition to time-independent constraints regarding aspects like scalability, resilience, and control plane communication delays, dynamically changing network conditions like traffic patterns or bandwidth demands need to be considered as well. Consequently, such dynamic environments call for a regular and fast recalculation of placements in order to adapt to the current situation in a timely manner. While an exhaustive evaluation of all possible solutions can be performed within a practically feasible time frame for small and medium-sized networks, such an approach is out of scope for large problem instances which have significantly higher time and memory requirements. Therefore, this work investigates a specialized heuristic, which takes into account a particular set of optimization objectives and returns solutions representing the possible trade-offs between them. Due to its low computation time and acceptable margin of error, this heuristic can be employed by automatic decision systems operating in dynamic environments.
• Investigating the Impact of Network Topology on the Processing Times of SDN Controllers. Metter, Christopher; Gebert, Steffen; Lange, Stanislav; Zinner, Thomas; Tran-Gia, Phuoc; Jarschel, Michael. In Seventh IFIP/IEEE International Workshop on Management of the Future Internet. Ottawa, Canada, 2015.
  • [ Abstract ]
  • [ BibTeX ]
  • [ EndNote ]
  • [ Download ]
  Software Defined Networking (SDN) introduces the concept of logically-centralized controllers in charge of managing the forwarding behavior of network elements. The new possibilities enabled through the centralization of control logic come with a certain risk: The controller might become a performance bottleneck. Therefore, ensuring sufficient controller performance is one of the crucial tasks prior to a successful SDN deployment. Furthermore, fine-grained traffic engineering, e.g., to achieve higher link utilization, results in a higher frequency of requests that are sent to the controller, which leads to an increased controller load. It is therefore important to analyze the capabilities of SDN controllers prior to deployment. This paper investigates two software implementations, the OpenDaylight and Ryu controllers. The control message throughput of different controllers has been studied several times already; however, it is not yet known what influence the number and topology of connected switches have. This paper investigates this influence in detail for a fat-tree data center topology and a WAN topology as well as 261 topologies with varying characteristics from the Internet Topology Zoo.

  @inproceedings{info3-inproceedings-2015-510, abstract = {Software Defined Networking (SDN) introduces the concept of logically-centralized controllers in charge of managing the forwarding behavior of network elements. The new possibilities enabled through the centralization of control logic come with a certain risk: The controller might become a performance bottleneck. Therefore, ensuring sufficient controller performance is one of the crucial tasks prior to a successful SDN deployment. Furthermore, fine-grained traffic engineering, e.g., to achieve higher link utilization, results in a higher frequency of requests that are sent to the controller, which leads to an increased controller load. It is therefore important to analyze the capabilities of SDN controllers prior to deployment. This paper investigates two software implementations, the OpenDaylight and Ryu controllers. The control message throughput of different controllers has been studied several times already; however, it is not yet known what influence the number and topology of connected switches have. This paper investigates this influence in detail for a fat-tree data center topology and a WAN topology as well as 261 topologies with varying characteristics from the Internet Topology Zoo.}, address = {Ottawa, Canada}, author = {Metter, Christopher and Gebert, Steffen and Lange, Stanislav and Zinner, Thomas and Tran-Gia, Phuoc and Jarschel, Michael}, booktitle = {Seventh IFIP/IEEE International Workshop on Management of the Future Internet}, keywords = {saser-siegfried}, month = 5, title = {Investigating the Impact of Network Topology on the Processing Times of SDN Controllers}, year = 2015 }

  %0 Conference Paper %1 info3-inproceedings-2015-510 %A Metter, Christopher %A Gebert, Steffen %A Lange, Stanislav %A Zinner, Thomas %A Tran-Gia, Phuoc %A Jarschel, Michael %B Seventh IFIP/IEEE International Workshop on Management of the Future Internet %C Ottawa, Canada %D 2015 %T Investigating the Impact of Network Topology on the Processing Times of SDN Controllers %X Software Defined Networking (SDN) introduces the concept of logically-centralized controllers in charge of managing the forwarding behavior of network elements. The new possibilities enabled through the centralization of control logic come with a certain risk: The controller might become a performance bottleneck. Therefore, ensuring sufficient controller performance is one of the crucial tasks prior to a successful SDN deployment. Furthermore, fine-grained traffic engineering, e.g., to achieve higher link utilization, results in a higher frequency of requests that are sent to the controller, which leads to an increased controller load. It is therefore important to analyze the capabilities of SDN controllers prior to deployment. This paper investigates two software implementations, the OpenDaylight and Ryu controllers. The control message throughput of different controllers has been studied several times already; however, it is not yet known what influence the number and topology of connected switches have. This paper investigates this influence in detail for a fat-tree data center topology and a WAN topology as well as 261 topologies with varying characteristics from the Internet Topology Zoo.
• Investigating Isolation between Virtual Networks in Case of Congestion for a Pronto 3290 Switch. Nguyen-Ngoc, Anh; Lange, Stanislav; Gebert, Steffen; Zinner, Thomas; Tran-Gia, Phuoc; Jarschel, Michael. In Workshop on Software-Defined Networking and Network Function Virtualization for Flexible Network Management (SDNFlex 2015). Cottbus, Germany, 2015.
  • [ Abstract ]
  • [ BibTeX ]
  • [ EndNote ]
  • [ Download ]
  Resource isolation between virtual networks is one of the key features of network virtualization. It is typically realized by configuring queues with specific rate guarantees on the egress ports of the network devices. The drawback of this architectural choice, however, is that traffic from several ingress ports may result in congestion on an egress port. Hence, the question arises to which extent isolation between virtual networks is realized in state-of-the-art hardware. This work aims at investigating whether congestion within one virtual network may affect the throughput performance of another virtual network. For that, measurements in a local testbed using a Pronto 3290 switch running an OpenFlow-enabling Pica8 firmware are performed.

  @inproceedings{info3-inproceedings-2015-508, abstract = {Resource isolation between virtual networks is one of the key features of network virtualization. It is typically realized by configuring queues with specific rate guarantees on the egress ports of the network devices. The drawback of this architectural choice, however, is that traffic from several ingress ports may result in congestion on an egress port. Hence, the question arises to which extent isolation between virtual networks is realized in state-of-the-art hardware. This work aims at investigating whether congestion within one virtual network may affect the throughput performance of another virtual network. For that, measurements in a local testbed using a Pronto 3290 switch running an OpenFlow-enabling Pica8 firmware are performed.}, address = {Cottbus, Germany}, author = {Nguyen-Ngoc, Anh and Lange, Stanislav and Gebert, Steffen and Zinner, Thomas and Tran-Gia, Phuoc and Jarschel, Michael}, booktitle = {Workshop on Software-Defined Networking and Network Function Virtualization for Flexible Network Management (SDNFlex 2015)}, keywords = {saser-siegfried}, month = 3, title = {Investigating Isolation between Virtual Networks in Case of Congestion for a Pronto 3290 Switch}, year = 2015 }

  %0 Conference Paper %1 info3-inproceedings-2015-508 %A Nguyen-Ngoc, Anh %A Lange, Stanislav %A Gebert, Steffen %A Zinner, Thomas %A Tran-Gia, Phuoc %A Jarschel, Michael %B Workshop on Software-Defined Networking and Network Function Virtualization for Flexible Network Management (SDNFlex 2015) %C Cottbus, Germany %D 2015 %T Investigating Isolation between Virtual Networks in Case of Congestion for a Pronto 3290 Switch %X Resource isolation between virtual networks is one of the key features of network virtualization. It is typically realized by configuring queues with specific rate guarantees on the egress ports of the network devices. The drawback of this architectural choice, however, is that traffic from several ingress ports may result in congestion on an egress port. Hence, the question arises to which extent isolation between virtual networks is realized in state-of-the-art hardware. This work aims at investigating whether congestion within one virtual network may affect the throughput performance of another virtual network. For that, measurements in a local testbed using a Pronto 3290 switch running an OpenFlow-enabling Pica8 firmware are performed.
• Heuristic Approaches to the Controller Placement Problem in Large Scale SDN Networks. Lange, Stanislav; Gebert, Steffen; Zinner, Thomas; Tran-Gia, Phuoc; Hock, David; Jarschel, Michael; Hoffmann, Marco. In IEEE Transactions on Network and Service Management - Special Issue on Efficient Management of SDN and NFV-based Systems, 12(1), pp. 4–17. 2015.
  • [ Abstract ]
  • [ BibTeX ]
  • [ EndNote ]
  • [ Download ]
  Software Defined Networking (SDN) marks a paradigm shift towards an externalized and logically centralized network control plane. A particularly important task in SDN architectures is that of controller placement, i.e., the positioning of a limited number of resources within a network in order to meet various requirements. These requirements range from latency constraints to failure tolerance and load balancing. In most scenarios, at least some of these objectives are competing, thus no single best placement is available and decision makers need to find a balanced trade-off. This work presents POCO, a framework for Pareto-based Optimal COntroller placement that provides operators with Pareto optimal placements with respect to different performance metrics. In its default configuration, POCO performs an exhaustive evaluation of all possible placements. While this is practically feasible for small and medium sized networks, realistic time and resource constraints call for an alternative in the context of large scale networks or dynamic networks whose properties change over time. For these scenarios, the POCO toolset is extended by a heuristic approach that is less accurate, but yields faster computation times. An evaluation of this heuristic is performed on a collection of real world network topologies from the Internet Topology Zoo. Utilizing a measure for quantifying the error introduced by the heuristic approach allows an analysis of the resulting trade-off between time and accuracy. Additionally, the proposed methods can be extended to solve similar virtual functions placement problems which appear in the context of Network Functions Virtualization (NFV).

  @article{info3-article-2015-131, abstract = {Software Defined Networking (SDN) marks a paradigm shift towards an externalized and logically centralized network control plane. A particularly important task in SDN architectures is that of controller placement, i.e., the positioning of a limited number of resources within a network in order to meet various requirements. These requirements range from latency constraints to failure tolerance and load balancing. In most scenarios, at least some of these objectives are competing, thus no single best placement is available and decision makers need to find a balanced trade-off. This work presents POCO, a framework for Pareto-based Optimal COntroller placement that provides operators with Pareto optimal placements with respect to different performance metrics. In its default configuration, POCO performs an exhaustive evaluation of all possible placements. While this is practically feasible for small and medium sized networks, realistic time and resource constraints call for an alternative in the context of large scale networks or dynamic networks whose properties change over time. For these scenarios, the POCO toolset is extended by a heuristic approach that is less accurate, but yields faster computation times. An evaluation of this heuristic is performed on a collection of real world network topologies from the Internet Topology Zoo. Utilizing a measure for quantifying the error introduced by the heuristic approach allows an analysis of the resulting trade-off between time and accuracy. Additionally, the proposed methods can be extended to solve similar virtual functions placement problems which appear in the context of Network Functions Virtualization (NFV).}, author = {Lange, Stanislav and Gebert, Steffen and Zinner, Thomas and Tran-Gia, Phuoc and Hock, David and Jarschel, Michael and Hoffmann, Marco}, journal = {IEEE Transactions on Network and Service Management - Special Issue on Efficient Management of SDN and NFV-based Systems}, keywords = {poco}, number = 1, pages = {4 - 17}, title = {Heuristic Approaches to the Controller Placement Problem in Large Scale SDN Networks}, volume = 12, year = 2015 }

  %0 Journal Article %1 info3-article-2015-131 %A Lange, Stanislav %A Gebert, Steffen %A Zinner, Thomas %A Tran-Gia, Phuoc %A Hock, David %A Jarschel, Michael %A Hoffmann, Marco %D 2015 %J IEEE Transactions on Network and Service Management - Special Issue on Efficient Management of SDN and NFV-based Systems %N 1 %P 4 - 17 %T Heuristic Approaches to the Controller Placement Problem in Large Scale SDN Networks %V 12 %X Software Defined Networking (SDN) marks a paradigm shift towards an externalized and logically centralized network control plane. A particularly important task in SDN architectures is that of controller placement, i.e., the positioning of a limited number of resources within a network in order to meet various requirements. These requirements range from latency constraints to failure tolerance and load balancing. In most scenarios, at least some of these objectives are competing, thus no single best placement is available and decision makers need to find a balanced trade-off. This work presents POCO, a framework for Pareto-based Optimal COntroller placement that provides operators with Pareto optimal placements with respect to different performance metrics. In its default configuration, POCO performs an exhaustive evaluation of all possible placements. While this is practically feasible for small and medium sized networks, realistic time and resource constraints call for an alternative in the context of large scale networks or dynamic networks whose properties change over time. For these scenarios, the POCO toolset is extended by a heuristic approach that is less accurate, but yields faster computation times. An evaluation of this heuristic is performed on a collection of real world network topologies from the Internet Topology Zoo. Utilizing a measure for quantifying the error introduced by the heuristic approach allows an analysis of the resulting trade-off between time and accuracy. Additionally, the proposed methods can be extended to solve similar virtual functions placement problems which appear in the context of Network Functions Virtualization (NFV).

• Evaluation of Video Quality Monitoring based on Pre-computed Frame Distortions. Klein, Dominik; Zinner, Thomas; Borchert, Kathrin; Lange, Stanislav; Singeorzan, Vlad; Schmid, Matthias. In 19th EUNICE Workshop on Advances in Communication Networking, Best Paper Award. Chemnitz, Germany, 2013.
  • [ Abstract ]
  • [ BibTeX ]
  • [ EndNote ]
  • [ Download ]
  A large fraction of the current Internet traffic is caused by video streaming. Due to the growing expectations of video consumers, monitoring video applications is getting more and more important for network and service providers. In a previous work, we proposed a video quality monitoring solution which utilizes the full reference SSIM metric to improve the monitoring in the network by distributing pre-computed distortion information induced by frame losses. To improve scalability, we introduced a less complex algorithm which infers the distortion for higher loss scenarios from single loss scenarios and inter-frame dependencies. In this work, we evaluate the accuracy of our algorithm by comparing it with the exact calculation of the SSIM metric for different frame loss scenarios. We further consider different high definition test video sequences and group of picture structures and investigate the influence on the accuracy of our proposed approximation.

  @inproceedings{info3-inproceedings-2013-474, abstract = {A large fraction of the current Internet traffic is caused by video streaming. Due to the growing expectations of video consumers, monitoring video applications is getting more and more important for network and service providers. In a previous work, we proposed a video quality monitoring solution which utilizes the full reference SSIM metric to improve the monitoring in the network by distributing pre-computed distortion information induced by frame losses. To improve scalability, we introduced a less complex algorithm which infers the distortion for higher loss scenarios from single loss scenarios and inter-frame dependencies. In this work, we evaluate the accuracy of our algorithm by comparing it with the exact calculation of the SSIM metric for different frame loss scenarios. We further consider different high definition test video sequences and group of picture structures and investigate the influence on the accuracy of our proposed approximation.}, address = {Chemnitz, Germany}, author = {Klein, Dominik and Zinner, Thomas and Borchert, Kathrin and Lange, Stanislav and Singeorzan, Vlad and Schmid, Matthias}, booktitle = {19th EUNICE Workshop on Advances in Communication Networking, Best Paper Award}, keywords = {video-quality}, month = 8, title = {Evaluation of Video Quality Monitoring based on Pre-computed Frame Distortions}, year = 2013 }

  %0 Conference Paper %1 info3-inproceedings-2013-474 %A Klein, Dominik %A Zinner, Thomas %A Borchert, Kathrin %A Lange, Stanislav %A Singeorzan, Vlad %A Schmid, Matthias %B 19th EUNICE Workshop on Advances in Communication Networking, Best Paper Award %C Chemnitz, Germany %D 2013 %T Evaluation of Video Quality Monitoring based on Pre-computed Frame Distortions %X A large fraction of the current Internet traffic is caused by video streaming. Due to the growing expectations of video consumers, monitoring video applications is getting more and more important for network and service providers. In a previous work, we proposed a video quality monitoring solution which utilizes the full reference SSIM metric to improve the monitoring in the network by distributing pre-computed distortion information induced by frame losses. To improve scalability, we introduced a less complex algorithm which infers the distortion for higher loss scenarios from single loss scenarios and inter-frame dependencies. In this work, we evaluate the accuracy of our algorithm by comparing it with the exact calculation of the SSIM metric for different frame loss scenarios. We further consider different high definition test video sequences and group of picture structures and investigate the influence on the accuracy of our proposed approximation.
• Video Quality Monitoring based on Precomputed Frame Distortions. Klein, Dominik; Zinner, Thomas; Lange, Stanislav; Singeorzan, Vlad; Schmid, Matthias. In IFIP/IEEE International Workshop on Quality of Experience Centric Management (QCMan). Ghent, Belgium, 2013.
  • [ Abstract ]
  • [ BibTeX ]
  • [ EndNote ]
  • [ Download ]
  In the past decade, video streaming has taken over a large part of the current Internet traffic and more and more TV broadcasters and network providers extend their portfolio of video streaming services. With the growing expectations of video consumers with respect to the service quality, monitoring is an important aspect for network providers to detect possible performance problems or high network load. In parallel, emerging technologies like software defined networking or network virtualization introduce support for specialized networks which allow enhanced functionality in the network. This development enables more sophisticated monitoring techniques in the specialized networks which use knowledge about the video content to better predict the service quality at consumers. In this work, we present a content-aware SSIM-based monitoring technique and compare it with the current state-of-the art which infers the service quality from the monitored packet loss. We further show how network conditions like packet loss or bursts influence the two different monitoring techniques.

  @inproceedings{info3-inproceedings-2013-464, abstract = {In the past decade, video streaming has taken over a large part of the current Internet traffic and more and more TV broadcasters and network providers extend their portfolio of video streaming services. With the growing expectations of video consumers with respect to the service quality, monitoring is an important aspect for network providers to detect possible performance problems or high network load. In parallel, emerging technologies like software defined networking or network virtualization introduce support for specialized networks which allow enhanced functionality in the network. This development enables more sophisticated monitoring techniques in the specialized networks which use knowledge about the video content to better predict the service quality at consumers. In this work, we present a content-aware SSIM-based monitoring technique and compare it with the current state-of-the art which infers the service quality from the monitored packet loss. We further show how network conditions like packet loss or bursts influence the two different monitoring techniques.}, address = {Ghent, Belgium}, author = {Klein, Dominik and Zinner, Thomas and Lange, Stanislav and Singeorzan, Vlad and Schmid, Matthias}, booktitle = {IFIP/IEEE International Workshop on Quality of Experience Centric Management (QCMan)}, keywords = {sys:hidden:inproceedings}, month = 5, title = {Video Quality Monitoring based on Precomputed Frame Distortions}, year = 2013 }

  %0 Conference Paper %1 info3-inproceedings-2013-464 %A Klein, Dominik %A Zinner, Thomas %A Lange, Stanislav %A Singeorzan, Vlad %A Schmid, Matthias %B IFIP/IEEE International Workshop on Quality of Experience Centric Management (QCMan) %C Ghent, Belgium %D 2013 %T Video Quality Monitoring based on Precomputed Frame Distortions %X In the past decade, video streaming has taken over a large part of the current Internet traffic and more and more TV broadcasters and network providers extend their portfolio of video streaming services. With the growing expectations of video consumers with respect to the service quality, monitoring is an important aspect for network providers to detect possible performance problems or high network load. In parallel, emerging technologies like software defined networking or network virtualization introduce support for specialized networks which allow enhanced functionality in the network. This development enables more sophisticated monitoring techniques in the specialized networks which use knowledge about the video content to better predict the service quality at consumers. In this work, we present a content-aware SSIM-based monitoring technique and compare it with the current state-of-the art which infers the service quality from the monitored packet loss. We further show how network conditions like packet loss or bursts influence the two different monitoring techniques.