Generating representative Web workloads for network and server performance evaluation
SIGMETRICS '98/PERFORMANCE '98 Proceedings of the 1998 ACM SIGMETRICS joint international conference on Measurement and modeling of computer systems
Congestion Control in Linux TCP
Proceedings of the FREENIX Track: 2002 USENIX Annual Technical Conference
Linux Network Architecture
Determining an appropriate sending rate over an underutilized network path
Computer Networks: The International Journal of Computer and Telecommunications Networking
Building a RCP (Rate Control Protocol) Test Network
HOTI '07 Proceedings of the 15th Annual IEEE Symposium on High-Performance Interconnects
The new web: characterizing AJAX traffic
PAM'08 Proceedings of the 9th international conference on Passive and active network measurement
Performance comparison of router assisted congestion control protocols: XCP vs. RCP
Proceedings of the 2nd International Conference on Simulation Tools and Techniques
Performance Evaluation of Fast Startup Congestion Control Schemes
NETWORKING '09 Proceedings of the 8th International IFIP-TC 6 Networking Conference
Performance evaluation of quick-start in low latency networks
CCNC'10 Proceedings of the 7th IEEE conference on Consumer communications and networking conference
Comparison of end-to-end and network-supported fast startup congestion control schemes
Computer Networks: The International Journal of Computer and Telecommunications Networking
FavorQueue: A parameterless active queue management to improve TCP traffic performance
Computer Networks: The International Journal of Computer and Telecommunications Networking
Hi-index | 0.00 |
Quick-Start is an experimental extension of the Transmission Control Protocol (TCP) that uses explicit router feedback to speed up best effort data transfers. With Quick-Start, TCP endpoints can request permission from the routers along the path to send at a higher rate than allowed by the default TCP congestion control, which avoids the time-consuming Slow-Start. However, since Quick-Start TCP requires modifications in the protocol stacks of end-systems and routers, realization complexity is a major concern. This paper studies Quick-Start with a new implementation in the Linux protocol stack. We first show that Quick-Start support can be added to a real stack with rather limited effort, without causing much processing overhead. Second, we perform measurements with Web applications and study the impact of important parameters. These experiments with real applications demonstrate that Quick-Start can significantly speed up data transfers, and they confirm the outcome of previous simulation efforts. Our results suggest that Quick-Start is a lightweight mechanism that could be very beneficial for broadband interactive applications in the future Internet.