2002
DOI: 10.1016/s1389-1286(02)00291-8
|View full text |Cite
|
Sign up to set email alerts
|

TCP-Real: receiver-oriented congestion control

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
2
1
1
1

Citation Types

0
33
0
1

Year Published

2005
2005
2012
2012

Publication Types

Select...
6
2

Relationship

2
6

Authors

Journals

citations
Cited by 57 publications
(34 citation statements)
references
References 24 publications
(44 reference statements)
0
33
0
1
Order By: Relevance
“…We believe that solutions like the one proposed in [35], in which the receiver controls the sender's cwnd, may be integrated into our final algorithm. Likewise, the work in [36] could be useful for improving the fairness of our mechanism by including the congestion window in the timeout computation at the sender.…”
Section: Discussionmentioning
confidence: 99%
“…We believe that solutions like the one proposed in [35], in which the receiver controls the sender's cwnd, may be integrated into our final algorithm. Likewise, the work in [36] could be useful for improving the fairness of our mechanism by including the congestion window in the timeout computation at the sender.…”
Section: Discussionmentioning
confidence: 99%
“…On the aspect of error classification [11], previously, we have proposed TCP-Real [12], which relies on a novel receiver-oriented and measurement-based congestion control mechanism to differentiate the error nature. If the packet loss is due to random wireless losses, the congestion window is not reduced and the transmission rate is maintained.…”
Section: A Proposalmentioning
confidence: 99%
“…Notably the lowest goodput of individual flow is less than 10% of the highest one, as demonstrated by worst-case fairness in Figure 8(b). It has already been shown in [17] that with DropTail routers and standard TCP, the main traffic consumed more bandwidth than the cross traffic, due to the fact that packets aggregated into link R1-R2 (see Figure 3) before entering the bottleneck R3 were more uniformly distributed in the time domain; therefore, having smaller probability to get dropped, compared with the bursty traffic of non-aggregated (i.e. connecting to R3 directly) cross-traffic flows.…”
Section: Network Behaviors Over Heterogeneous (Wired/wireless) Networkmentioning
confidence: 67%