An instrumentation and analytics framework for optimal and robust NFV deployment
暂无分享,去创建一个
This article details a novel approach to the fine-tuning of carrier-grade virtualized network function deployments for both performance optimization and diagnostic purposes, using embedded instrumentation with an analytics framework. The work presented in this article is the output of a co-lab established between BT and Intel Labs Europe to investigate network-function-virtualization-related problems faced by traditional telecom operators. Results from comprehensive testing of virtual CDN and WAN acceleration use cases are presented. These results led to the development of a number of insights that will help network operators optimize the deployment of NFV on standard high volume servers. Significantly, it was found that the default configuration for some use cases resulted in suboptimal resource allocation and consumption. Consequently, opportunities exist for carriers to fine-tune their NFV deployments from both the technical and economic perspectives using approaches such as embedded instrumentation.
[1] Raj Jain,et al. Network virtualization and software defined networking for cloud computing: a survey , 2013, IEEE Communications Magazine.
[2] Brendan Gregg,et al. Systems Performance: Enterprise and the Cloud , 2013 .