K. Varadhan, D. Estrin, and S. Floyd. Impact of network dynamics on end-to-end protocols: Case studies in tcp and reliable multicast. Technical Report USC-CS-TR 98-672, University of Southern California, Information Sciences Institute, April 1998.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....and a drop into fast recovery at high throughputs, due to loss of packets being transmitted between nodes, or to later interleaving of packets and acknowledgements in flight along both original and new routes as routing information is propagated. Such transient effects are discussed further in [13]. It is unlikely that any broadband satellite constellation will be a true IP packet switching network based on the IP packet structure. Instead, the need for management of frequency allocation in the terminal uplink and downlink dictates the need for MAC with a fixed frame size. A number of ....

K. Varadhan, D. Estrin, and S. Floyd, "Impact of Network Dynamics on End-to-End Protocols: Case Studies in TCP and Reliable Multicast," Tech. rep. 98-672, Dept. of Comp. Sci., USC, Mar. 1998.

....that can only be done in the TCP operation While we do not claim to have answers to these questions for all types of scenarios, our work attempts to shed some light in this area. Before concluding the introductory discussion, we like to draw attention to the recent work of Varadhan et al. [35] who have addressed the impact of network dynamics on TCP s behavior. They have examined several TCP flavors on a single three core node network topology that ran the distance vector routing protocol and have considered a single link failure. In particular, they have found that network dynamic ....

K. Varadhan, D. Estrin and S. Floyd, "Impact of Network Dynamics on End-to-End Protocols: Case Studies in TCP and Reliable Multicast," Technical report USC-CS-TR 98-672, University of Southern California, April 1998.

....be informed of the loss and required to assist in recovering from the loss. Active retransmission techniques can be used when the application can tolerate larger end to end delays. A widely deployed multicast scheme based on the retransmission of lost packets is Scalable Reliable Multicast (SRM) [VEF98]. Audio Loss Repair Sender Based Receiver Based Active Passive Insertion Interpolation Regeneration Figure 2.1 Taxonomy of Audio Loss Repair Techniques Receiver based repair techniques are also called error concealment. These techniques can be initiated by the receiver of an audio stream ....

K. Varadhan, D. Estrin, S. Floyd, Impact of Network Dynamics on End-to-End Protocols: Case Studies in TCP and Reliable Multicast, In the Proceedings of the IEEE Symposium on Computers and Communications, July 1998.

....can give much better performance for large simulations, but that, in some cases, produces slightly different results. To determine if the differences caused by abstraction would affect the end results of a real protocol study, we examine SRM timer behavior (previously explored in [50] 63] [66]) with both detailed and abstract simulations. 5.1.1. SRM Mechanism SRM (Scalable Reliable Multicast) is a reliable multicast transport protocol. Each member detects losses individually and issues multicast requests for retransmission per loss. Any member who successfully receives the packets may ....

....from A. After a while, member B records T sendA when sending another session message to member A, and, similarly, member A records T rcvB upon receiving the session message. The one way distance is estimated by: T rcvB T sendA T rcvA T sendB ) 2 Details are in [50] Refinements are in [63] [66]. 5.1.2. Applying Session Multicast The three common metrics used to estimate the effectiveness of SRM are recovery delay, number of duplicate requests, and number of duplicate repairs. Recovery delay is the time duration from loss detection to recovery. Number of duplicate requests is the number ....

Kannan Varadhan, Deborah Estrin, and Sally Floyd, Impact of Network Dynamics on Endto -End Protocols: Case Studies in Reliable Multicast, Submitted for Review to the Third IEEE symposium on Computers and Communications, http://netweb.usc.edu/vint/papers/dynamics.ps, University of Southern California, August, 1997

....describe a performance sensitive routing algorithm used in the early ARPANET [MRR80] and Khanna et al. KZ88] discuss the behavior of this algorithm under varying degrees of load. Varadhan et al. present a simulation study of the effect of path changes on the performance of transport protocols [VEF98] They show that small path changes during a TCP session can lead to significant reordering and a consequent reduction in performance. Finally, Francis et al. explore the possibility of using end to end measurements to construct maps of the minimum Internet propagation delay between hosts [FJP ....

Kannan Varadhan, Deborah Estrin, and Sally Floyd. Impact of Network Dynamics on End-to-End Protocols: Case Studies in TCP and Reliable Multicast. Technical Report USC-CS-TR 98-672, University of Southern California, Information Sciences Institute, April 1998.

....describe a performance sensitive routing algorithm used in the early ARPANET [MRR80] and Khanna et al. KZ88] discuss the behavior of this algorithm under varying degrees of load. Varadhan et al. present a simulation study of the effect of path changes on the performance of transport protocols [VEF98] They show that small path changes during a TCP session can lead to significant reordering and a consequent reduction in performance. Finally, Francis et al. explore the possibility of using end to end measurements to construct maps of the propagation delay along different paths [FJP 99] ....

Kanna Varadhan, Deborah Estrin, and Sally Floyd. Impact of network dynamics on end-to-end protocols: Case studies in tcp and reliable multicast. Technical Report USC-CS-TR 98-672, University of Southern California, Information Sciences Institute, April 1998.

No context found.

K. Varadhan, D. Estrin, and S. Floyd, "Impact of network dynamics on end-to-end protocols: Case studies in reliable multicast," in Proc. 3rd IEEE Symp. Computers and Communications , 1998, pp. 147--153.

....delay matrix, D. Mapping of delay matrix into complex topologies is out of scope of this document. This topology generator is probably representative of a standard tool for topology generation used in networking research. Using GT ITM we have covered most topologies used in several SRM studies [18] [19] We faced difficulties when choosing the lossy link for the random topologies in order to maximize the number of responses. This is an example of the difficulties networking researchers face when trying to stress networking protocols in an ad hoc way. SRM response timer values are ....

K. Varadhan, D. Estrin, and S. Floyd. Impact of Network Dynamics on End-to-End Protocols: Case Studies in Reliable Multicast. ISCC98, 1998.

....responses. Whereas random topologies simulation generated almost 10 responses in the worst case. 11 The topology generator is probably representative of a standard tool for topology generation used in networking research. Using GT ITM we have covered most topologies used in several SRM studies [18] [19] 12 We faced di#culties when choosing the lossy link for the random topologies in order to maximize the number of responses. This is an example of the di#culties networking researchers face when trying to stress networking protocols in an ad hoc way. 13 SRM response timer values are ....

K. Varadhan, D. Estrin, and S. Floyd. Impact of Network Dynamics on End-to-End Protocols: Case Studies in Reliable Multicast. ISCC98, 1998.

....12 . The results of the simulation are shown in Figure 6. The number of responses triggered for 10 The topology generator is probably representative of a standard tool for topology generation used in networking research. Using GT ITM we have covered most topologies used in several SRM studies [18] [19] 11 We faced difficulties when choosing the lossy link for the random topologies in order to maximize the number of responses. This is an example of the difficulties networking researchers face when trying to stress networking protocols in an ad hoc way. 12 SRM response timer values are ....

K. Varadhan, D. Estrin, and S. Floyd. Impact of Network Dynamics on End-to-End Protocols: Case Studies in Reliable Multicast. ISCC98, 1998.

....are indirect. Broader use of nam suggests that visualization is more than just a tool for fancy demos, but that it can substantially ease protocol debugging and help understand dynamic behavior. Because of these reasons, a growing number of researchers have used nam in their work and papers [16, 9]. Acknowledgments Marylou Orayani made substantial contributions to nam as part of her work at Berkeley in 1995 and 1996. Nam has also bene ted from an enthusiastic VINT and ns user community. We would like to thank especially Elan Amir, Lee Breslau, Kevin Fall, Sally Floyd, Ahmed Helmy, Polly ....

Varadhan, K., Estrin, S., and Floyd, S. Impact of network dynamics on end-to-end protocols: Case studies in reliable multicast. In Proceedings of the International Symposium on Computers and Communications (Aug. 1998). http:// www.isi.edu/~kannan/papers/iscc98.ps.gz.

....receives the repair from another node. The recovery delay observed by any node is a function of the topology. Therefore, one possible definition that we can use is the average recovery delay for a loss; this is the average of the delays observed by each of the nodes that experienced that loss [19]. The average recovery delay is also affected by the topology, which node initiates the request and which other nodes assists in the repair. These metrics can have high variability. 1 In particular, note that we are using dense mode multicast, based on periodic flood and prune. In our topology, ....

....has a greater impact on the loss recovery delays that are of interest to us. Therefore, in this paper, we restrict our characterisation of protocol performance to the number of request messages sent for each loss. The complete characterisation of the protocol based on the repair messages is in [19]. Figure 3 shows these plots of request messages for fixed and adaptive timers. In order to show all of the different points in the distribution, we plot the distribution as jittered dots. From Fig 0 0.5 1 1.5 2 2.5 3 3.5 0 20 40 60 80 100 120 time (a) # of requests sent in fixed timer ....

[Article contains additional citation context not shown here]

K. Varadhan, S. Estrin, and S. Floyd. Impact of network dynamics on end-to-end protocols: Case studies in TCP and reliable multicast. Technical Report USC CS TR 98-672, University of Southern California, Department of Computer Science, Mar. 1998. http://www.isi.edu/ kannan/ papers/ impact-tr.ps.gz.

....can give much better performance for large simulations, but that in some cases it produces slightly different results. To determine if the differences caused by abstraction would affect the end results of a real protocol study, this section examines SRM timer behavior (previously explored in [26, 31 34]) with both detailed and abstract simulations. 5.1. SRM Mechanism SRM (Scalable Reliable Multicast) 26] is a reliable multicast transport protocol. Each member detects losses individually and issues multicast requests for retransmission per loss. Any member who successfully receives the packets ....

Kannan Varadhan, Deborah Estrin, and Sally Floyd, Impact of Network Dynamics on End-to-End Protocols: Case Studies in Reliable Multicast, Submitted for Review to the Third IEEE symposium on Computers and Communications, http://netweb.usc.edu/vint/papers/dynamics.ps, University of Southern California, August, 1997

No context found.

K. Varadhan, D. Estrin, and S. Floyd. Impact of network dynamics on end-to-end protocols: Case studies in tcp and reliable multicast. Technical Report USC-CS-TR 98-672, University of Southern California, Information Sciences Institute, April 1998.

No context found.

K. Varadhan, D. Estrin, and S. Floyd, "Impact of Network Dynamics on End-to-End Protocols: Case Studies in TCP and Reliable Multicast," USC/Information Science Institute, California, Technical Report USC-CS-TR 98-672, April 1998.

No context found.

K. Varadhan, D. Estrin and S. Floyd, "Impact of Network Dynamics on End-to-End Protocols: Case Studies in TCP and Reliable Multicast," Technical report USC-CS-TR 98-672, USC, April 1998.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC