دانلود رایگان مقاله کدهای جاسازی شبکه مجازی برای حفاظت از شبکه درجه مخابراتی

Abstract

The decoupling of software from hardware by means of virtualization presents us with a unique opportunity to perform on-the-fly network deployment and reconfiguration. Virtual Machines could be instantiated and virtual links can connect these machines to form end-to-end virtualized networks on top of a physical network infrastructure. Virtual Network Embedding (VNE) algorithms could be used to map such virtual network on a physical network infrastructure. However, present VNE algorithms do not consider overall network protection. This is critical for a Telecom operator that needs to realize 99.999% network availability. Moreover, VNE is an NP-hard problem. In this paper, we propose a heuristics VNE aimed at protecting Telecom operator sites. Our objective is to develop advance counter-measures in the form of telco-grade redundancy to avoid large scale network failures such as the one observed during the tsunami in Japan in 2011. In designing our VNE algorithm, we distinguish server nodes from switching nodes as a server function cannot be embedded on a switching node. We also choose not to employ path splitting which is difficult to implement in real network operations. Along with detail modeling of our proposal, we also evaluate our scheme which shows its efficiency in realizing operators' valuable infrastructure protection while maintaining lower resource consumption.

7. Conclusions

In this paper, we have proposed and evaluated a heuristic algo- rithm which realizes VNE for 1 + 1 site protection to meet Telco- grade network protection and service availability requirements. In approaching this issue, we do not relax the problem by using path splitting as that is unrealistic in operational networks. We also dis- tinguish between server nodes and switching nodes, as a server node cannot be mapped on a switching node. Our algorithm achieves sound correlation between node and 968 link embedding, consumes less bandwidth and provides higher success rate than any conventional approach. Evaluation results also show that our algorithm performance in finding out a VNE solution is close to the theoretical ceiling. In our future work, we will address VN node mapping over multiple sites in order to fur- ther optimize site resources and backup path optimization in order to further reduce the consumption of link resources. In addition, we intend to further analyze the performance of the proposed VNE 975 algorithm in failure scenarios focusing of service recovery latency.