Energy-Efficient, QoS-Aware Packet Scheduling in High-Speed Networks

In current commercial routers, faster network processor units (NPUs) in line cards (LCs) can significantly improve network QoS performance. However, this improvement may come at a high cost of energy consumption. In this paper, we propose and investigate two classes of QoS-aware DVFS-based packet scheduling schemes. The first class uses queue length (QL) to control execution rates in line cards, whereas the second uses link utilization to achieve the same purpose. Excessive reduction in execution rate to save energy, however, may result in a sharp increase in network delay. To address this challenge, different frequency scaling strategies are proposed, and the performance of their associated schedulers is investigated. The simulation results show that both QL-aware and Load-aware DVFS have potential for significant energy saving in high-speed networks, with acceptable delay performance. Furthermore, a simulation study is carried out to compare the performance of the proposed schemes to similar schemes described in the literature for different network environments and traffic loads. The results show that the mQ̅LA scheme achieves the best results, with performance gains of up to 9.5% energy saving, while meeting the QoS performance of the supported applications.

[1]  Balazs Sonkoly,et al.  Multi-platform performance evaluation of digital fountain based transport , 2014, 2014 Science and Information Conference.

[2]  Xue Liu,et al.  TailCon: Power-Minimizing Tail Percentile Control of Response Time in Server Clusters , 2012, 2012 IEEE 31st Symposium on Reliable Distributed Systems.

[3]  Luca Valcarenghi,et al.  Sleep Mode for Energy Saving PONs: Advantages and Drawbacks , 2009, 2009 IEEE Globecom Workshops.

[4]  Edmond A. Jonckheere,et al.  On the predictability of data network traffic , 2003, Proceedings of the 2003 American Control Conference, 2003..

[5]  Wes Simpson Video Over IP: A Practical Guide to Technology and Applications , 2005 .

[6]  Christina Hattingh,et al.  End-to-end qos network design , 2005 .

[7]  Sergiu Nedevschi,et al.  Reducing Network Energy Consumption via Sleeping and Rate-Adaptation , 2008, NSDI.

[8]  Lachlan L. H. Andrew,et al.  Common TCP Evaluation Suite , 2009 .

[9]  Stephan Bohacek,et al.  Dynamic Modeling of Internet Traffic for Intrusion Detection , 2007, EURASIP J. Adv. Signal Process..

[10]  Nian-Feng Tzeng,et al.  Energy-efficient scheme for multiprocessor-based router linecards , 2006, International Symposium on Applications and the Internet (SAINT'06).

[11]  Juan Li,et al.  An overview of energy efficiency techniques in cluster computing systems , 2013, Cluster Computing.

[12]  Marco E. T. Gerards,et al.  Algorithmic power management - Energy minimisation under real-time constraints , 2014 .

[13]  Juanita Ellis,et al.  Voice, Video, and Data Network Convergence: Architecture and Design, From VoIP to Wireless , 2003 .

[14]  Stephen J. Wright,et al.  Power Awareness in Network Design and Routing , 2008, IEEE INFOCOM 2008 - The 27th Conference on Computer Communications.

[15]  Marina Thottan,et al.  Adapting router buffers for energy efficiency , 2011, CoNEXT '11.

[16]  Rodney S. Tucker,et al.  Modeling Energy Consumption in High-Capacity Routers and Switches , 2014, IEEE Journal on Selected Areas in Communications.