The Naval Expeditionary Warfare Internetworking project ("NEW" for short) was funded by ONR at the University of California, Santa Cruz (UCSC). This three-year project ended in 2002.
At UCSC, this project was part of the research carried out within the Computer Communication Research Group (CCRG) of the Baskin School of Engineering.
The principal investigator of this project was J.J. Garcia-Luna-Aceves.
The NEW Internetworking project seeks to develop new communication protocols for the NEW environmemnt. The innovations we intend to introduce with this project are:
Our approach consisted of advancing the state of the art in the following areas: We defined a new model of internetworking, and determine how it can deliver better functionality than the IP Internet model. The problems that we addressed to accomplish this include the following:
We started the definition of our federated internetworking architecture. The salient features of the architecture are its use of anycasting services. A detailed description of the new architecture is planned to be ready by the end of the summer and a paper will be submitted for publication at that time.
We completed the first design of TCP-Santa Cruz (or TCP-SC). TCP-SC is a a new implementation of TCP that can be implemented as a TCP option by utilizing the extra 40 bytes available in the options field. TCP-SC detects not only the initial stages of congestion, but can also identify the direction of congestion, i.e., it determines whether congestion is developing in the forward path and then isolates the forward throughput from events such as congestion that may occur on the reverse path. The direction of congestion is determined by estimating the relative delay that one packet experiences with respect to another; this relative delay is the foundation of our congestion control algorithm. Our approach is significantly different from rate-controlled congestion control approaches, e.g., TCP-Vegas, as well as those which use increasing round-trip time (RTT) as a primary indication of congestion, in that we do not use RTT estimates in any way for congestion control. This represents a fundamental improvement over the latter approaches, because round-trip time measurements are inherently misleading in that they cannot distinguish between variation due to increases or decreases in the forward or reverse paths of the connection.
TCP-SC provides a better error-recovery strategy than Reno and Tahoe by providing a mechanism to perform RTT estimates for every packet transmitted, including retransmissions. This eliminates the need for Karn's algorithm and does not need to use any timer-backoff strategies that can lead to long idle periods on the links. In the absence of congestion in the reverse path or asymmetric links (i.e., if all transmitted acknowledgment packets are received), TCP-SC also provides a mechanism to avoid retransmission by timeout when there are several losses per window.
Several simulation experiments were run to show the performance improvements of TCP-SC over Reno, Tahoe and Vegas.
The definition of our federated internetworking architecture continued.
Recent research on TCP has focussed on the problems associated with TCP performance in the presence of wireless links and ways to improve its performance. During this reporting period, we extended our work on TCP-SC by introducing an extension to it that improves TCP performance over lossy wireless links. TCP has no mechanism to differentiate random losses on the wireless link from congestion, and therefore treats all losses as congestive. We developed a simple method in which TCP-SC is able to differentiate these random losses, thereby avoiding the rate-halving approach taken by standard TCP whenever any loss is detected. We have compared the performance of our protocol against TCP Reno using simulation experiments runnign in ns2, which show that TCP-SC achieves higher throughput and lower end-to-end delay than today's TCP.
In addition to our work on TCP, we addressed the use of Forward Error Correction (FEC) techniques in reliable multipoint communication. Mor especifically, we studied group loss probabilities of FEC codes in shared loss multicast communication. Most other research to date has studied independent loss in multicast trees. It has extended non-FEC, single packet analysis to the FEC realm. We revised this form of analysis with state equations specific to C(n,k) codes, where any k out of n packets may decode an entire group. Rather than analyzing the number of transmissions of a particular packet, we studied an FEC group as a whole. Using recursive equations, we found the cumulative distribution function that all leaf nodes in a shared loss tree successfully decode a C(n,k) FEC transmission group. Our analysis also yields a method for computing the probability mass function that r leaf nodes successfully decode a transmission group on the first transmission. We may also compute for a particular leaf node the expected number of packets received on successful decode and the expected number of missing packets on decode failure. Our analytic results closely match simulation runs that we performed over a variety of configurations.
To analyze retransmissions, we developed a set of theorems to calculate the packet correlation between leaf nodes in a full regular tree. We were able to compute the probability mass function that all leaf nodes hold the same m packets in common, given that the source transmitted an ordered sequence of n packets.
We applied our analysis to several multicast topologies. We evaluated packet reception and ACK/NAK generation for the first transmission of an FEC group. Retransmissions are studied both with and without subcasting. We also compared retransmissions with and without FEC-enhancement.
The following is the list of published papers describing our research results in this project. A more complete list of CCRG publications, including PDF format, can be found here.