Improvements at the physical network layer have enabled technologies such as 10 gigabit Ethernet. Single core end-systems are unable to fully utilize these networks, due to limited clock cycles. Using a multi-core architecture is one method which increases the number of available cycles, and thus allow networks to be fully utilized. However, using these systems creates a new set of challenges for network protocol processing, for example, deciding how best to utilize many cores for high network performance.This paper examines different ways the cores of a multi-core system can be utilized, and, by experimentation, we show that in an eight core system deciding which cores to use is important. In one test, there was a 40% discrepancy in CPU utilization depending on which cores were used. This discrepancy results from the resources each core shares, an example being the multi-hierarchy CPU caches, and to which bus the processors are connected.

This work proposes new retransmission and backoff strategies for network-wide broadcasting in multi-hop wireless networks. A comparative analysis is presented between existing algorithms as well as the ones proposed herein. Simulation experiments and analysis are used throughout this work to study or demonstrate the properties and performance of specific strategies as well as other properties or results of a more general nature. Several topics not considered in previous work are also studied. The broadcasting strategies are evaluated with respect to their impact on routing protocols that rely on flooding to perform path discovery, and research is conducted in designing schemes that maximize the route lifetime. Different backoff strategies are proposed and their performance is examined.

The authors propose an architecture where clients can make advance reservations through agents. For each routing domain in the network there will be an agent responsible for admission control on behalf of the routers in the domain. Requests involving several routing domains are forwarded for admission control with agents along the path for the requested service. Agents maintain hard reservation state using a reliable protocol for agent intercommunication. Agents start allocating resources for advance reservations in the routers by setting up forwarding state shortly before resources are needed for packet fop warding. Resources are made available for advance reservations by means of rejecting further immediate requests and ultimately by preempting some immediate reservations. They have shown that the risk of preemption can be kept very low. Thus, agents can set up packet classifiers and schedulers in their routers, allowing routers to get on with their main task, packet forwarding.

For some time, the networking community has assumed that it is impossible to do IP routing lookups in software fast enough to support gigabit speeds. IP routing lookups must find the routing entry with the longest matching prefix, a task that has been thought to require hardware support at lookup frequencies of millions per second.We present a forwarding table data structure designed for quick routing lookups. Forwarding tables are small enough to fit in the cache of a conventional general purpose processor. With the table in cache, a 200 MHz Pentium Pro or a 333 MHz Alpha 21164 can perform a few million lookups per second. This means that it is feasible to do a full routing lookup for each IP packet at gigabit speeds without special hardware.The forwarding tables are very small, a large routing table with 40,000 routing entries can be compacted to a forwarding table of 150-160 Kbytes. A lookup typically requires less than 100 instructions on an Alpha, using eight memory references accessing a total of 14 bytes.

Wireless i~; becoming a pop•ular way to connect mobile computers to the Internet and other networks.The bandwidth of wireless links will probably always be limited d1lf to properties of the physical medium and regulatory limits on the n~;e of frcqncncies for radio communication.Therefore.it is necessary for network protocols to utilize the available bandwidth efficiently.If enders of IP packc/s arc growing and the bandwidth required for transmitting headers is increasing.With the coming of I Pvfj the addre8.ssize in creases from 4 to 16 bytes and the basic IP head a increases from 20 to 4 0 bytes.Monovcr, most mobility schemes tunnel packets addressed to mobile hosts by adding an extra JP header o1• eJ:lm routing information, typically increasing the .size of TCP/IPv4 headers to 60 bytes and TCP/!Pvfi headers to 100 bytes.In this paper, we provide new header compres~;ion schemes for UDP/IP and TCP/IP protocols.IVc show how to reduce the size of UDP /IP headers by an order of magnitude, down to four to five bytes.Our method worl.'s Ol'el' simplex links, lossy links, multi-access links, and supporls multicast communication.IVe also shou.: how to generali.~ethe mo::;l commonly used method for header compres~;ion for 'J'CP/IP!•.{, dereloped by liau Jacobson, to !Pv6 and nwltiple JP headtrs.The resulting scheme nnfortun at ely reduces TCP throughput ova los~;y link$ due to unfavorable interaction with TCP's congestion control mechanisms.However, by addin_q two•This work was supported by grants from the c..,nt.re for Dist.anceSpanning Technology (COT).Lul<!t•, Sweden, and Ericsson Radio Systems AB.

As an experiment in protocol optimizations, the authors undertook to improve the performance of a stateless protocol, namely the user datagram protocol (UDP) in the 4.3 BSD Unix kernel. The authors describe the successful optimizations that were done, along with measurements that shows a UDP performance improvement of between 25-35% on CISC and RISC systems, and overall kernal improvement of between 12% and 18%.>