INET framework extensions for TCP Vegas and TCP Westwood

Author(s):  
Maria Fernandez ◽  
Carlos Calafate ◽  
Juan-Carlos Cano ◽  
Pietro Manzoni
Keyword(s):  
2010 ◽  
Vol 30 (9) ◽  
pp. 2485-2487
Author(s):  
Bin WANG ◽  
Yuan-yan CHEN ◽  
Yu HU ◽  
Wei FENG

2018 ◽  
Vol 0 (0) ◽  
Author(s):  
Reza Poorzare ◽  
Siamak Abedidarabad

AbstractThere is a misunderstanding in Optical Burst Switching (OBS) networks about the congestion status in the network that can cause a reduction in the performance of the networks. OBS networks are bufferless in their nature so when a burst drop happens in the network it can be because of the congestion or contention in the network but TCP cannot distinguish it is due to the congestion or contention. TCP wrongly decreases the congestion window size (cwnd) and causes significant reduction of the network performance. In this paper we are trying to employ a new algorithm by using fuzzy logic and some thresholds to divide the network into several areas then we can solve the problem. This new scheme can help us to distinguish a burst drop is because of the congestion or a burst contention in the network. Extensive simulative studies show that the proposed algorithm outperforms TCP Vegas in terms of throughput and packet delivery count.


2018 ◽  
Vol 7 (2.12) ◽  
pp. 231
Author(s):  
Sukantkishoro Bisoy ◽  
Pradeep Kumar Mallick ◽  
Anjana Mishra

Conservative nature of Vegas creates less opportunity to get fair share of bandwidth then Reno in wired network. On the other hand, aggressive nature of Reno helps to achieve more share of bandwidth. Both Reno and Vegas assumes that congestion occurs in the forward rather than in reverse path. In asymmetric network the path characteristics of forward and backward is different.In this work, we propose a network model and analyzed the Inter-protocol fairness between TCP Reno and TCP Vegas with some queue management techniques such as Droptail and random early detection (RED) in asymmetric network where the forward and backward path has different characteristics. The simulation experiment results using NS2 indicates that use of RED can achieve better fairness than Droptail in asymmetric network.  


Sign in / Sign up

Export Citation Format

Share Document