TY - GEN
T1 - Virtualized congestion control
AU - Cronkite-Ratcliff, Bryce
AU - Bergman, Aran
AU - Vargaftik, Shay
AU - Ravi, Madhusudhan
AU - McKeown, Nick
AU - Abraham, Ittai
AU - Keslassy, Isaac
N1 - Publisher Copyright:
© 2016 ACM.
PY - 2016/8/22
Y1 - 2016/8/22
N2 - New congestion control algorithms are rapidly improving datacenters by reducing latency, overcoming incast, increasing throughput and improving fairness. Ideally, the operating system in every server and virtual machine is updated to support new congestion control algorithms. However, legacy applications often cannot be upgraded to a new operating system version, which means the advances are off-limits to them. Worse, as we show, legacy applications can be squeezed out, which in the worst case prevents the entire network from adopting new algorithms. Our goal is to make it easy to deploy new and improved congestion control algorithms into multitenant datacenters, without having to worry about TCP-friendliness with non-participating virtual machines. This paper presents a solution we call virtualized congestion control. The datacenter owner may introduce a new congestion control algorithm in the hypervisors. Internally, the hypervisors translate between the new congestion control algorithm and the old legacy congestion control, allowing legacy applications to enjoy the benefits of the new algorithm. We have implemented proof-of-concept systems for virtualized congestion control in the Linux kernel and in VMware's ESXi hypervisor, achieving improved fairness, performance, and control over guest bandwidth allocations.
AB - New congestion control algorithms are rapidly improving datacenters by reducing latency, overcoming incast, increasing throughput and improving fairness. Ideally, the operating system in every server and virtual machine is updated to support new congestion control algorithms. However, legacy applications often cannot be upgraded to a new operating system version, which means the advances are off-limits to them. Worse, as we show, legacy applications can be squeezed out, which in the worst case prevents the entire network from adopting new algorithms. Our goal is to make it easy to deploy new and improved congestion control algorithms into multitenant datacenters, without having to worry about TCP-friendliness with non-participating virtual machines. This paper presents a solution we call virtualized congestion control. The datacenter owner may introduce a new congestion control algorithm in the hypervisors. Internally, the hypervisors translate between the new congestion control algorithm and the old legacy congestion control, allowing legacy applications to enjoy the benefits of the new algorithm. We have implemented proof-of-concept systems for virtualized congestion control in the Linux kernel and in VMware's ESXi hypervisor, achieving improved fairness, performance, and control over guest bandwidth allocations.
KW - Algorithmic virtualization
KW - DCTCP
KW - Datacenters
KW - ECN
KW - Hypervisors
KW - TCP
KW - Virtualized congestion control
UR - http://www.scopus.com/inward/record.url?scp=84986537103&partnerID=8YFLogxK
U2 - 10.1145/2934872.2934889
DO - 10.1145/2934872.2934889
M3 - ???researchoutput.researchoutputtypes.contributiontobookanthology.conference???
AN - SCOPUS:84986537103
T3 - SIGCOMM 2016 - Proceedings of the 2016 ACM Conference on Special Interest Group on Data Communication
SP - 230
EP - 243
BT - SIGCOMM 2016 - Proceedings of the 2016 ACM Conference on Special Interest Group on Data Communication
T2 - 2016 ACM Conference on Special Interest Group on Data Communication, SIGCOMM 2016
Y2 - 22 August 2016 through 26 August 2016
ER -