The following research has all been supported by NSF and the Office of Naval Research. Please click here for more recent papers.

IP Lookups

"Fast Scalable Routing Lookups". Proceedings of ACM Sigcomm, Sep 97. Describes a new approach to solving the IP lookup problem that is exponentially faster than previous schemes, has been patented and used by industry.

"Faster IP Lookups using Controlled Prefix Expansion". Proceedings of ACM Sigmetrics,Sep 98 and ACM TOCS 99. Another approach to solving the IP lookup problem problem that has been patented and used by industry.

"IP Lookups using Multiway and Multicolumn Search". IP Lookups using a new technique called Multicolumn Binary Search. Journal version of paper that appeared in Infocom 98.

" Packet Classification using Tuple Space Search". New algorithms for fast packet filtering for firewalls and other applications. SIGCOMM 99.

"Fast and Scalable Layer 4 Switching". Describes algorithms for fast packet classification for firewalls and resource allocation. We have code available that takes existing firewall databases and provides much faster search than current router firewall implementations even for large databases.

Fair Queuing

"Leap Forward Virtual Clock: An O(log log N) Queuing Scheme with Guaranteed Delays and Throughput fairness". Proceedings of Infocom 97 Shows the first O(log log N) Fair Queuing Scheme for Internet Routers. Fair Queuing schemes are an important component of the next generation internet to accomodate video and audio. Previous schemes took O(log N), where N is the number of concurrent conversations.

"Efficient Fair Queueing using Deficit Round Robin" SIGCOMM 95 and ACM/IEEE Trans Networking. An efficient O(1) substitute for the O(log(n) sorting bottleneck in DKS fair queueing. Provides throughput fairness only and reasonable (but not optimal) delay bounds. Implemented by Cisco and Microsoft.

"Randomized Token Buckets: reducing the buffers in a multiplexor". Proceedings of Nossdav, May 97. A new approach to traffic shaping based on a randomized version of leaky buckets. 


Techniques to Overcome Other Bottlenecks in Protocol Design

"Reliable and Efficient Striping". SIGCOMM 96. Shows new connections between fair queuing and load balancing protocols using a time-reversal argument. Uses this to develop and implement practical load balancing protocols that ensure FIFO delivery. Implemented in Cisco's IOS 11.3.

"Trading Packet Headers for Packet Processing" Describes additions to headers to speed up packet processing. SIGCOMM 95. Among 5 papers nominated for (and accepted by) ACM/IEEE Trans Networking. Describes a scheme called threaded indexing that is essentially identical to a recent industry standard called Tag Switching.

"Hashed and Hierarchical Timing Wheels: data structures to efficiently implement a timer facility" IEEE Trans Networking, Dec 97. Revised from a 1984 SOSP paper; used in X kernel, Siemens and DEC OS's etc.

"Redesigning the BSD Callout Facility" Shows how our earlier ideas on timing wheels can be used to efficiently reimplement the BSD timer facilities. In Software Practice and Experience. Code available from Adam Costello's home page.

"Reconsidering Fragmentation and Reassembly". PODC 98 Fast reassembly, new schemes for partial reassembly, techniques to cope with instability. 


Papers on Protocol Design

"An Error Control Scheme for Large Scale Multicast Applications". Scalable reliable multicast with minimal router assistance. Full version based on Infocom 98 paper, with source code.

"Efficient and Reliable Hop-by Flow Control" Appeared in SIGCOMM 94 and IEEE JSAC 95. New flow control techniques for ATM networks using credit based flow control in which credits for each VC can be shared in a safe manner.

"The Pessimism behin Optimistic Simulation" PADS 94 and Inf Proc. Letters 1995. A new Global Virtual Time (GVT) algorithm for distributed simulation derived from a termination detection protocol due to Misra.

"Bounding the Unbounded", Infocom 94. Using reset protocols to convert unbounded counter protocols to bounded counter protocols.

"A Tradeoff between Safety and Liveness for Randomized Coordinated Attack" PODC 92 and Information and Computation, July 96, shows the (in)effectiveness of randomization in curing the impossibility of coordinated attack. Chapter 5 in Lynch's book on Distributed Computing is devoted to these results. 


Robust Protocol Design and Self-stabilization

"Fault Span of Crash Failures". Shows crash faults can drive systems to arbitrarily muddled states. Journal version accepted to JACM based on papers in PODC 1996 and PODC 1997. PODC 96 paper won best student paper award. Justifies the need for self-stabilizing protocols.

"Self-stabilization by counter flushing". Self-stabilization using a fresh counter value. Final journal version based on paper in PODC 1994.

"Self-stabilization by Local Checking and Correction". When and how Local Checking and Correction suffices. Final journal paper based on FOCS 91 paper and my Ph.D. thesis.

"Self-stabilization by Window Washing". PODC 1996. Shows how to make window protocols self-stabilizing and applies this idea to flow control and ATM broadcast protocols.

"Self-stabilization by Local Checking and Network Reset", WDAG - 94. Completes the local checking saga by showing that any locally checkable protocol can be made self-stabilizing using a reset protocol. Proof details can be found in my thesis.

"Self-stabilizing End-to-End Communication" , expanded version of a protocol described rather sketchily in FOCS-91. Journal of High Speed Networks, 1996.

"Compositional Proofs of Self-stabilizing Protocols" . 3rd Workshop on Self-stabilization, 1997.

"Local Checking and Tree correction" 2nd Workshop on Self-stabilizing Protocols and Chicago Journal of CS.

"Understanding Dijkstra's Self-stabilizing Protocols" . Shows how to understand two of the algorithms in Dijkstra's seminal paper using our general methods.

"Summary of Ph.D. Thesis on Self-stabilization" . A summary of my Ph.D Thesis that should be easier to read than the thesis itself.

"Distributed Constraint Satisfaction" (in DCS94, with Arora and Gouda), on self-stabilization via constraint satisfaction. (please send mail to anish@cis.ohio-state.edu for a postscript file) 


Prepared by George Varghese, varghese@ccrc.wustl.edu
Last Modified Oct 99.