ABSTRACT
Developing optimized distributed protocols for large-scale networks is a challenging problem due to scalability and stability concerns. Scalability concerns can be naturally addressed by interpreting distributed protocols as a non-cooperative game of local protocol components attempting to maximize their individual utilities. One of the difficulties in implementing this approach is developing adaptive algorithms capable of learning of the expected utilities and adjusting the corresponding control actions for the purpose of approaching the solution to the corresponding game, and thus optimization of the global system performance. It is known that the best response by each component to its expected utility may result in unstable behavior and deterioration of the overall performance. On an example of cross-layer optimization of a TCP/IP network, this paper discusses the possibility of avoiding these undesirable effects by allowing the control actions occasionally deviate from their best response values. Using simulations, the paper suggests that (a) sufficient level of randomness in route selection improves the network performance by eliminating the route flapping instability, (b) the optimal level of randomness keeps the network within the stability region in close proximity to the border of this region, and (c) it may be possible to optimize the network performance by adjusting the level of randomness.
- A. G. Ganek and T. A. Corbi, "The dawning of the autonomic computing era," IBM Systems J., Vol. 42, No 1, 2003, pp.5--18. Google ScholarDigital Library
- D. Fudenberg and D. K. Levine, The theory of learning in games, The MIT Press, 1999.Google Scholar
- C. Huitema, Routing in the Internet, Prentice Hall, 2000. Google ScholarDigital Library
- D. Bertsekas and R. Gallager, Data Networks, Prentice-Hall, New Jersey, 1992. Google ScholarDigital Library
- J. Wang, L. Li, S. H. Low, and J. C. Doyle, "Cross-layer optimization in TCP/IP networks," IEEE/ACM Trans. on Networking, to appear, available at http://netlab.caltech.edu/. Google ScholarDigital Library
- A. Elwalid, C. Jin, S. Low and J. Widjaja, "MATE: MPLS Adaptive Traffic Engineering," in Proceedings of Infocom'2001, Anchorage, Alaska, April 2001.Google Scholar
- D. P. Bertsekas, "Dynamic behavior of shortest path routing algorithm for communication networks," IEEE Trans. On Automatic Control, pp. 60--74, Feb. 1982.Google ScholarCross Ref
- S. H. Low and P. Varaiya, "Dynamic behavior of a class of adaptive routing protocols (IGRP)," Proc. Of Infocom'93, pp. 610--616, March 1993.Google Scholar
- Z. Wang, Y. Wang, and L. Zhang, "Internet traffic engineering without full mesh overlaying," in Proceedings of Infocom'2001, Anchorage, Alaska, April 2001.Google Scholar
- A. Sridharan, R. Guerin, and Cristophe Diot, "Achieving Near Optimal Traffic Engineering Solution for Current OSPF/IS-IS Networks," in Proceedings of Infocom'2003.Google Scholar
- F. P. Kelly, A. K. Maulloo, and D. H. K. Tan, "The rate control for communication networks: shadow prices, proportional fairness and stability," Journal of the Operational Research Society, pp. 237--252, vol. 409, 1998.Google Scholar
- T. M. Cover and J. A. Thomas, Elements of Information Theory, Wiley, 1991. Google ScholarDigital Library
- V. Marbukh, "Network provisioning as a game against nature," IEEE International Communications Conference (ICC 2003), Anchorage, AK, 2003.Google Scholar
- V. Marbukh, "On Shortest random walks under adversarial uncertainty", Fortieth Annual Allerton Conference on Communication, Control, and Computing, Monticello, Illinois, 2002.Google Scholar
- O. Borchert and R. Rouil, "The GMPLS lightwave agile switching simulator -- an overview," http://www.antd.nist.gov/glass.Google Scholar
Index Terms
- Decentralized control of large-scale networks as a game with local interactions: cross-layer TCP/IP optimization
Recommendations
Decentralized Guaranteed Cost Control for Uncertain Large-Scale Systems Using Delayed Feedback: LMI Optimization Approach
In this paper, we propose a design method of guaranteed cost controllers for uncertain large-scale systems with time delays in subsystem interconnections using delayed feedback. Using the Lyapunov method, a linear matrix inequality (LMI) optimization ...
Dynamics of hot-potato routing in IP networks
Despite the architectural separation between intradomain and interdomain routing in the Internet, intradomain protocols do influence the path-selection process in the Border Gateway Protocol (BGP). When choosing between multiple equally-good BGP routes, ...
Domain Constrained Multicast: A New Approach for IP Multicast Routing
One of major reasons why IP multicast has not been well deployed is the complexity of IP multicast routing. Since existing IP multicast routing protocols have been designed independently of IP unicast routing protocols, a router must maintain routing ...
Comments