| Eggert, L., Heidemann, J., and Touch, J., "Effects of Ensemble TCP," ACM Computer Comm. Review, January 2000. |
....results comparing the session fair algorithms with each other and with the original connection fair algorithm are presented in Section V. Implementation issues are briefly outlined in Section VI, and we conclude in Section VII. II. RELATED WORK Previous work, such as Webmux [13] Ensemble TCP [14], and Integrated Congestion Manager [15] looks at the problem of fairness of resource allocation to multiple connections between the same client and server pair. In our work, we extend this problem to the case of multiple servers sending to a single client. The challenge in this case is that each ....
L. Eggert, J. Heidemann, and J. Touch, "Effects of Ensemble-TCP," ACM Computer Communication Review, vol. 30, no. 1, pp. 15--29, January 2000.
....new. T TCP [19] and TCP Fast Start [20] improve TCP performance for small data transfers by caching and reusing connection state. Henderson et al. [21] have investigated the effects of TCP algorithms on bandwidth allocation between concurrent connections. Congestion Manager [10] and Ensemble TCP [22] share state information between connections to a remote host and allow the aggregate bandwidth to be divided among state sharing connections by a priority mechanism. TCP Nice [23] adjusts TCP s congestion control algorithm to ensure that background TCP flows have lower priority than regular ....
Lars Eggert, John Heidemann, and Joe Touch, "Effects of Ensemble-TCP," ACM Computer Communication Review, vol. 30, no. 1, pp. 15--29, Jan. 2000.
....Hari Balakrishnan et al. has shown in [22] that flows can learn from each other and share information about the congestion state along the network path, which it terms as shared state learning. They use this as a basis for their congestion manager [22] Similarly, Lars et al. present Ensemble TCP [33] that provides control sharing of state information between TCP connections and aggregation amoung number of connections in parallel. Using per aggregate schedulers, they show that TCP performance improves by 4 75 when compared to TCP Reno, while web download improves by about 17 61 using ....
L. Eggert, J. Heidemann and J. Touch, "Effects of Ensemble-TCP" ACM Computer Communication Review, 30(1), pages 15-29, January, 2000
.... startup and recovery from multiple TCP packet losses have also been proposed [2] 3] 5] 9] 13] 17] 30] We believe that our context aware TCP IP approach complements many of these prior approaches, including connection state caching, rate based pacing, SACK TCP [13] Ensemble TCP [12], and RED [16] Barford and Crovella [8] discuss the critical path analysis of TCP connections, in an attempt to pinpoint the causes of the delays incurred during Web document transfers, However, this analysis is primarily from a network trafc analysis viewpoint, rather than a TCP control ....
L. Eggert, J. Heidemann, and J. Touch, "Effects of Ensemble-TCP", ACM Computer Communication Review, Vol. 30, No. 1, pp. 15-29, January 2000.
....on the target page are transferred in parallel. In this case, the web page retrieval latency is always determined by the largest object S2 S3 S4 S5 R0 R1 3Mb 10Mb 20ms 10ms C1 10Mb 10ms C3 C5 C2 . Fig. 10. A Network with Two Tiers of Clients. embedded [35] and can be simply estimated as its transmission latency. Sequential retrieval: The number of allowed concurrent connection is 1: the web objects are transferred sequentially (similar to HTTP 1.1 with persistent connection) In this case, the web page retrieval latency can be estimated as the ....
....latency of all web objects embedded. We believe that these two case studies give us some hints about what the real latency of web page retrieval is likely to be. Further more, they may also show the bounds under certain circumstances. We start with 5 representative types of web pages [35] listed in Table V. We estimate the estimated page retrieval latency with S SFD under both parallel and sequential retrieval schemes and compare the result with DT case. The absolute values and improvement (in percentage) are summarized in Table VI (positive value means reduction in retrieval ....
Lars Eggert, John Heidemann, and Joe Touch, "Effects of ensemble-TCP," ACM Computer Communication Review, vol. 30, no. 1, pp. 15--29, Jan. 2000.
No context found.
Eggert, L., Heidemann, J., and Touch, J., "Effects of Ensemble TCP," ACM Computer Comm. Review, January 2000.
No context found.
L. Eggert, J. Heidemann and J. Touch, "Effects of Ensemble-TCP" ACM Computer Communication Review, 30(1), pages 15-29, January, 2000
No context found.
EGGERT, L., HEIDEMANN, J., AND TOUCH, J. Effects of ensemble-tcp.
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