دانلود رایگان مقاله برنامه ریزی بهینه شبکه ارائه محتوا مجازی تحت تقاضای ترافیک نا مطمئن

Abstract

Content Delivery Networks (CDNs) have been identified as one of the relevant use cases where the emerging paradigm of Network Functions Virtualization (NFV) will likely be beneficial. In fact, virtualization fosters flexibility, since on-demand resource allocation of virtual CDN nodes can accommodate sudden traffic demand changes. However, there are cases where physical appliances should still be preferred, therefore we envision a mixed architecture in between these two solutions, capable to exploit the advantages of both of them. Motivated by these reasons, in this paper we formulate a two-stage stochastic planning model that can be used by CDN operators to compute the optimal long-term network planning decision, deploying physical CDN appliances in the network and/or leasing resources for virtual CDN nodes in data centers. Key findings demonstrate that for a large range of pricing options and traffic profiles, NFV can significantly save network costs spent by the operator to provide the content distribution service.

5. Conclusion

In this paper we tackled the stochastic planning problem for content delivery to study potential benefits that Network Functions Virtualization can provide for content distribution purposes. We considered a mixed architecture where both physical as well as virtual CDN nodes can be used by a CDN owner to implement the content distribution service. The owner performs the planning choice for the physical CDN infrastructure on a long-term time schedule, possessing only a stochastic estimate of future traffic demands. Our study shows that a mixed solution where both virtual an physical CDN nodes are used can dramatically reduce the overall costs sustained by the operator to purchase and operate the distribution infrastructure. In particular, we observed that gains can be up to 65% when considering the cheapest vCDN price. Our contribution is also to formulate efficient solution algorithms for the two-stage stochastic planning problem that can scale to realistic topology sizes. Rather than solving the deterministic equivalent problem in the extensive form, our proposed L-shaped algorithm and the greedy heuristic can efficiently find a solution, saving up to 95% of time compared to the MILP solver.