Browsing by Subject "Telecommunication--Traffic--Management"
Now showing 1 - 5 of 5
- Results Per Page
- Sort Options
Item Flow-size based differentiation to enhance user perceived performance on networks supporting best-effort traffic(2001) Yang, Shan-chieh; De Veciana, GustavoItem Measurement-based traffic management for QoS guarantee in multi-service networks(2001-05) Sang, Aimin; Li, San-QiItem Network congestion control(2001-08) Yang, Yang; Lam, Simon S., 1947-In a shared network such as the Internet, end systems should react to congestion by adapting their transmission rates to avoid congestion collapse and to keep network utilization high. The robustness of the current Internet is due in large part to the end-to-end congestion control mechanisms of TCP. Although TCP congestion control is appropriate for applications such as bulk data transfers, many new applications would find TCP’s behavior of halving the sending rate of a flow to be too severe a response to a single congestion indication, as it can noticeably reduce the flow’s user-perceived quality. However, since the dominant Internet traffic is TCP-based, it is important that new congestion control schemes be TCP-friendly. By this, we mean that the sending rate of a non-TCP flow should be approximately the same as that of a TCP flow under the same conditions of round-trip time and packet loss. The first problem I investigate is a more general version of AIMD (GAIMD) than what is implemented in TCP; specifically, the sender’s window size is increased by alpha if there is no packet loss in a round-trip time, and the window size is decreased to beta of the current value if there is a triple-duplicate loss indication, where alpha and beta are parameters. Using the relationship between the sending rate of GAIMD and the two parameters, I derive a simple relationship between alpha and beta for a GAIMD flow to be TCP-friendly. This relationship offers a wide selection of possible values for alpha and beta to achieve desired transient behaviors. I then investigate the fairness, smoothness, responsiveness, and aggressiveness of TCP, GAIMD and two other representative TCP-friendly congestion control protocols. The properties of these protocols are evaluated both analytically and via simulation by studying their responses to three network changes. Considering the inherent fluctuations in a stationary network environment, I define three types of sending rate variations, and derive an analytical expression for the CoV for each of the four protocols. I also study protocol responsiveness and aggressiveness by evaluating their responses to a step increase of network congestion and a step increase of available bandwidth. The third problem I investigate is the congestion control issues in multicast environments. A multicast session may have a large number of receivers with heterogeneous reception capacities determined by the fairness requirement of a network or by device capacity. To accommodate this heterogeneity, various multi-rate schemes, based upon the use of layering or replication, have been proposed. For a general class of receiver utility functions, I show that there exists an optimal partition that is ordered, which gives rise to efficient algorithms to find an optimal partition based upon dynamic programming. I also show algorithms to efficiently determine the optimal sending rate of each partitioned group. Furthermore, for several typical distributions of receivers’ capacities, I show that the majority of the benefit of a multi-rate scheme can be gained by using a small number of layers (or groups).Item Network tomography based on flow level measurements(2004) Arifler, Dogu; Evans, Brian L.; De Veciana, GustavoThe primary aim of network tomography is to infer properties of networks from network traffic measurements. Internet traffic mainly consists of flows of packets that belong to World Wide Web accesses, file transfers, and e-mail, whose transmissions are mediated via the Transmission Control Protocol (TCP). TCP flow records, or non-intrusive, flow level measurements, can be collected by the state-of-the-art networking equipment. In this dissertation, I develop a methodology to process TCP flow records to analyze throughput correlations among TCP flow classes. Throughputs of TCP flows that share resources in the network are correlated. These correlations can be used to infer resource sharing in the Internet. My proposal for using flow level measurements to infer network properties differs signifi- cantly from previous network tomography research that has employed packet level measurements for making inferences. In this work, I develop a sampling strategy for random processes (flow class throughputs) whose samples are taken when the processes are active at the sampling instant. The samples are used to estimate a flow class throughput correlation matrix. Factor analysis is then employed to investigate the correlation structure of TCP flow throughputs and to explore which TCP flow classes might share congested resources. A number of empirical studies are conducted to evaluate the effect of filtering out small or large sized flows on correlation estimates. Bootstrap methods are coupled with exploratory factor analysis to make inferential statements about resource sharing. The applicability of the methods to real datasets is also validated. Possible applications of the methodology introduced in this dissertation include network monitoring and root cause analysis of poor performance. The methods will have a potential impact on service providers who wish to analyze network performance using flow level measurements. The methodology may also be integrated into the design of future network monitoring equipment and software to perform an off-line evaluation of the congestion status of networks.Item Robust congestion control for IP multicast(2003) Gorinsky, Sergey; Vin, Harrick M.IP multicast is a network service for scalable distribution of data to multiple receivers. Traditional protocols for multicast congestion control rely on trust: each party is assumed to follow guidelines for fair bandwidth sharing. However, with the growth and commercialization of the Internet, the assumption of universal trust is no longer tenable. In this dissertation, we consider a relaxed model where receivers are untrustworthy and can misbehave to acquire an unfairly high bandwidth at the expense of competing traffic. Our experiments with existing multicast protocols show that each of the evaluated protocols is vulnerable to receiver misbehavior. To take the first step towards robust multicast designs for distrusted environments, we focus on the class of feedback-free protocols where receivers provide no feedback to the sender and control congestion by regulating their subscription levels in the multi-group session. Unfortunately, the mechanism of group subscription offers a misbehaving receiver an opportunity to inflate its subscription level. Such inflated subscription attacks pose a major threat to fairness of bandwidth allocation. This dissertation is the first to solve the problem of inflated subscription. The presented designs rely on an insight that the ability of a receiver to access a multicast group should be tied with the congestion status of the receiver. First, we address individual attacks where a receiver inflates its subscription with no assistance from other receivers. Our solution guards access to multicast groups with dynamic keys and consists of two independent components: DELTA (Distribution of ELigibility To Access) – a novel method for in-band distribution of group keys to receivers that are eligible to access the groups according to the congestion control protocol, and SIGMA (Secure Internet Group Management Architecture) – a generic architecture for key-based group access at edge routers. DELTA and SIGMA require only minimal generic changes in the edge routers, do not alter the core of the network, and introduce no auxiliary servers. Then, we extend the design to protect multicast congestion control against inflated subscription of colluding receivers. To illustrate that integration with DELTA and SIGMA makes multicast protocols robust to inflated subscription and preserves other congestion control properties, we derive and evaluate robust adaptations of RLM and FLID-DL protocols.