| Mark Allman. Improving TCP Performance Over Satellite Channels. Master's thesis, Ohio University, June 1997. |
.... (generally more than three seconds) to reach the full recommended window size in satellite links, which can waste satellite bandwidth in the slow start phase of TCP evolution [18] The steady state behavior of regular TCP cannot fully utilize the bandwidth provided by T1 satellite channels [19] Satellite Transport Protocol (STP) The key to Satellite Transport Protocol (STP) 17] is saving the reverse satellite channel bandwidth to overcome the asymmetry of satellite networks. First, it suggests the use of a split TCP scheme, i.e. the TCP connection is split at the gateway ....
M. Allman, "Improving TCP Performance over Satellite Channels, " Master's thesis, College of Engineering and Technology, Ohio University, June 1997.
....the segments, the segments will be queued for later processing. If a segment arrives at a router that has no memory left to queue the segment, the segment will be discarded. Therefore, it is important for TCP to be able to adapt its sending rate to the network conditions to avoid segment loss [Allm97]. Congestion control algorithms of TCP attempt to prevent congestive collapse by detecting congestion and reducing the transmission rate accordingly. While these algorithms are very important, they can also have a negative impact on the performance of TCP over wireless networks [CaIf94] The four ....
....have also investigated compressing the TCP and IP headers in an eoeort to reduce the overhead, leaving more bandwidth for the data [DENP97] 7.2 Larger Initial Window One possible solution to TCP problems is to use a larger initial sliding window. One study on this theory can be found in [Allm97]. The researchers used a modied FTP program called XFTP with several enhancements for satellite links, among which is a larger initial sliding window. The modications made the slow start algorithm more aggressive. The algorithm modications and experimental results showed an increase in bandwidth ....
Allman, M., Improving TCP Performance over Satellite Channels. Master of Science Thesis, Ohio University, OCLC # 37604261, June 1997.
....suggests an initial window size of roughly 4 Kbytes 14 , instead of one segment. This change would only apply at the beginning of the connection, in the rst round trip time, i.e, if TCP re enters Slow Start after a retransmit timeout, the congestion window is re initialized to one segment. In [1] larger initial windows are shown to improve TCP performance over satellite channels. This study shows a 27 throughput increase, using the 4 segment initial window suggested by Floyd, for 30 Kbyte FTP transfers, and a 180 throughput increase, using a 32 segment initial window. However, these ....
....the negative eoeects of delayed ACKs, which reduce by half the ACKs AEow rate, hence the congestion window growth. TCP s window increase algorithm should take into account the number of new segments covered by the ACK rather than the number of ACKs received. This modication was also evaluated in [1]. The study shows a throughput improvement of up to 15 (depending upon transfer size) as compared to standard TCP. The Slow Start threshold estimation Hoe [13] suggests that the initial Slow Start threshold value 15 be determined using both the the RTT and the available bandwidth of the ....
[Article contains additional citation context not shown here]
Mark Allman. Improving TCP Performance Over Satellite Channels. Master's thesis, Ohio, June 1997.
....sending application is injecting an inappropriate amount of data into the network. This causes the queuing buffers of the intermediate router to be flooded with excessive data, resulting in a packet loss. If the sending rate of the application is not controlled, it can lead to a congestion collapse[All97]. The various congestion control algorithms[APS99] used by TCP are : 7 Slow Start and Congestion Avoidance The slow start and congestion avoidance algorithms[APS99] allow TCP to increase the data rate without overwhelming the intermediate routers. To implement these algorithms, two variables ....
Mark Allman. Improving TCP Performance Over Satellite Channels. Master's thesis, Ohio University, August 1997.
....this option, the sender can get enough information from the receiver on the state of the correctly received packets and thus it is able to retransmit in one round trip time multiple lost segments. Simulation based performance results of TCP SACK over low and high delay paths were documented in [2, 20] and showed that TCP SACK can significantly improve the network performance when compared with earlier TCP implementations. Matthew Mathis and Jamshid Mahdavi in [31] proposed the Forward Acknowledgment (FACK) algorithm to reinforce TCP SACK for higher performance. FACK improves TCP congestion ....
....Performance Networks Ever since Slow Start and the Congestion Control Algorithms were first proposed, there have been constant modifications to TCP and to the congestion control algorithms to get better performance. A lot of research has been done on improving TCP performance on Satellite links [23, 2, 27]. The original TCP standard limits the TCP receive window to 65535 bytes. 11 Without the large windows extension proposed to TCP, the maximum throughput of a TCP connection is bounded by the round trip time [32] Throughput### # Receive Buffer Size Round Trip Time Without TCP large windows, ....
[Article contains additional citation context not shown here]
Mark Allman. Improving TCP Performance over Satellite Channels, Master of Science Thesis, Ohio University, 1997.
No context found.
Mark Allman. Improving TCP Performance Over Satellite Channels. Master's thesis, Ohio University, June 1997.
No context found.
Mark Allman. Improving TCP Performance Over Satellite Channels. Master's thesis, Ohio University, June 1997.
No context found.
Mark Allman. Improving TCP Performance Over Satellite Channels. Master's thesis, Ohio University, June 1997.
No context found.
Mark Allman. Improving TCP Performance Over Satellite Channels. Master's thesis, Ohio University, June 1997.
....section surveys simulations and experiments that explore the effect of larger initial windows on TCP connections. The first set of experiments explores performance over satellite links. Larger initial windows have been shown to improve the performance of TCP connections over satellite channels [All97b]. In this study, an initial window of four segments (512 byte MSS) resulted in throughput improvements of up to 30 (depending upon transfer size) KAGT98] shows that the use of larger initial windows results in a decrease in transfer time in HTTP tests over the ACTS satellite system. A study ....
Mark Allman. Improving TCP Performance Over Satellite Channels. Master's thesis, Ohio University, June 1997.
....which is also running NetBSD 1.2.1. The queue size in the routers have been adjusted to provide adequate buffering (roughly twice the size of TCP s advertised window in these experiments) The queue size used has been shown to prevent segment loss in previous bulk transfer experiments over ACTS [All97b] 4.2 Experimental Setup The experimental setup consists of running an HTTP browser on the LeRC side of the network, and an HTTP server on the OU side. The client application was Netscape s Navigator, version 3.01, for the HTTP 1.0 tests and W3C s WebBot 6 for the HTTP 1.1 tests. The WWW ....
Mark Allman. Improving TCP Performance Over Satellite Channels. Master's thesis, Ohio University, June 1997.
....transmit both segments was available when the transfer started. 100 1000 10000 100000 1e 06 1e 07 0 0.5 1 1.5 2 2.5 3 3.5 4 Time (seconds) Satellite Network Terrestrial Network Figure 1: Data transferred as a function of time over satellite and terrestrial network paths. Figure 1 from [All97] illustrates the low utilization of a satellite network during slow start, as compared to a network with a terrestrial delay (80 ms in this model) Just before 4 seconds into the transfer over the satellite link the slow start phase completes. During that same amount of time, the terrestrial ....
....a given file, rather than one connection as specified in [PR85] This multiplied TCP s aggressiveness by the number of TCP connections being utilized. The syntax and semantics of the extensions to FTP are outlined in [AO97] The ACTS experiments involving xftp are outlined in [AOK95, AKO96, All97] Figure 2 shows the throughput of a 5 MB transfer as a function of the number of parallel data connections used to transfer the file over an ACTS T1 link. Each connection used an advertised (maximum) window of 24 KB which 2 yields throughput of approximately 44,000 bytes second, as outlined ....
[Article contains additional citation context not shown here]
Mark Allman. Improving TCP Performance Over Satellite Channels. Master's thesis, Ohio University, June 1997.
....Data to Send Another question that researchers must consider is how much data should be transmitted and what applications should be used to send the data. Transfers that are small compared to the delay bandwidth product of the network have been shown to underutilize the available bandwidth (in [All97b] for example) and therefore experiments with short transfers only show a portion of TCP s behavior. On the other hand, very long transfers show TCP s steady state behavior quite nicely, but the startup behavior is lost. While both short transfers and long transfers provide solid data points ....
Mark Allman. Improving TCP Performance Over Satellite Channels. Master's thesis, Ohio University, June 1997.
....the available bandwidth of a network path has been the subject of a number of recent research studies. During slow start TCP underutilizes the capacity of the network path, especially if the round trip time (RTT) between the sender and the receiver is large (such as in satellite networks) All97] Several strategies have been proposed to mitigate the impact slow start has on performance. Hoe96] observes that the initial slow start period is often terminated by a large number of lost segments, hurting TCP s performance. Hoe96] suggests setting TCP s slow start threshold (ssthresh) state ....
Mark Allman. Improving TCP Performance Over Satellite Channels. Master's thesis, Ohio University, June 1997.
No context found.
Mark Allman. Improving TCP Performance Over Satellite Channels. Master's thesis, Ohio University, June 1997.
No context found.
Allman, M., Improving TCP Performance Over Satellite Channels, . 1997, Ohio University.
No context found.
Mark Allman. Improving TCP Performance Over Satellite Channels. Master's Thesis, Ohio University, June 1997.
No context found.
Allman, M. Improving TCP Performance Over Satellite Channels. Master's thesis, Ohio University, June 1997.
No context found.
Allman, Mark: "Improving TCP Performance Over Satellite Channels." Master's thesis, Ohio University, June 1997.
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