Opportunistic flooding to improve TCP transmit performance in virtualized clouds

Abstract

Virtualization is a key technology that powers cloud computing platforms such as Amazon EC2. Virtual machine (VM) consolidation, where multiple VMs share a physical host, has seen rapid adoption in practice with increasingly large number of VMs per machine and per CPU core. Our investigations, however, suggest that the increasing degree of VM consolidation has serious negative effects on the VMs' TCP transport performance. As multiple VMs share a given CPU, the scheduling latencies, which can be in the order of tens of milliseconds, substantially increase the typically submillisecond round-trip times (RTTs) for TCP connections in a datacenter, causing significant degradation in throughput. In this paper, we propose a light-weight solution called vFlood that (a) allows a TCP sender VM to opportunistically flood the driver domain in the same host, and (b) offloads the VM's TCP congestion control function to the driver domain in order to mask the effects of VM consolidation. Our evaluation of a vFlood prototype on Xen suggests that vFlood substantially improves TCP transmit throughput with minimal per-packet CPU overhead. Further, our application-level evaluation using Apache Olio, a web 2.0 cloud application, indicates a 33% improvement in the number of operations per second.

Understanding the network level performance of virtualization solutions

“…in the case of data exchange within a data center. To tackle this problem the authors in [3] propose to reinforce the cooperation between the hypervisor and the transport layer of the VMs. In [4], the authors pinpoint the factors that limit the maximum achievable rate under Xen.…”

Networking in a virtualized environment: The TCP case

“…Indeed, in a native or virtualized systems, heavily loading the CPU can lead to bad performance for a sender of network traffic. For instance, in [2], it has been reported that if a VM needs to transfer data while the remaining VMs execute computationally intensive CPU operations, the first VM can experience short breaks during its data transfer. The same phenomenon has been reported in [1] and [3].…”

“…This is achieved thanks to a combination of virtual NICs offered to the virtual servers and virtual switches that enable to interconnect the virtual NICs to the physical ones. To understand the impact of virtualization on the network performance, the main strategy in the studies carried out by the scientific community consists in stressing the virtual machines sharing the physical server with heavy CPU requirements, or heavy I/O operation to the hard disks [1], [2], [3].…”

Multihoming aware optimization mechanism

“…Evaluation considered a communication intensive service [Gamage et al, 2011]. As such, a file transfer service was employed and included transmission of data in different sizes: CD (750MB) considers ISO images that could be stored in compact discs.…”