Optimization of content delivery networks server placement

Content Delivery Networks (CDN) became in the last years an important element of the Internet architecture. CDN have many benefits, particularly their ability to enhance QoS perceived by users and the reduction of the cost of content delivery. Content is stored in many servers placed strategically close to potential users. In this paper, we consider the problem of optimizing CDN server placement. First, we propose a new modeling approach of the Internet and the CDNs based on Stochastic Geometry. Different point processes on the plane are used to represent the topology of the Internet (Interconnection of Autonomous Systems) and the placement of CDN servers. A cost function reflecting the server costs and the quality of service perceived by the users is then proposed. The optimization of the cost function allows us to quantify the optimal number of servers and their optimal placement. The impact of the different parameters on the system is also evaluated. We first analyze a system based on only one type of servers. Then, we analyze the more general case where different server classes and object placement polices are considered. For the last case, we are able to compare different policies for content placement that can take into account the popularity of different content elements, and to optimize the design of these policies.

[1]  François Baccelli,et al.  Stochastic geometry and architecture of communication networks , 1997, Telecommun. Syst..

[2]  Li Fan,et al.  Web caching and Zipf-like distributions: evidence and implications , 1999, IEEE INFOCOM '99. Conference on Computer Communications. Proceedings. Eighteenth Annual Joint Conference of the IEEE Computer and Communications Societies. The Future is Now (Cat. No.99CH36320).

[3]  Lili Qiu,et al.  On the placement of Web server replicas , 2001, Proceedings IEEE INFOCOM 2001. Conference on Computer Communications. Twentieth Annual Joint Conference of the IEEE Computer and Communications Society (Cat. No.01CH37213).

[4]  Sudipto Guha,et al.  Hierarchical placement and network design problems , 2000, Proceedings 41st Annual Symposium on Foundations of Computer Science.

[5]  Mark Crovella,et al.  Characteristics of WWW Client-based Traces , 1995 .

[6]  Arun Venkataramani,et al.  Bandwidth constrained placement in a WAN , 2001, PODC '01.

[7]  F. Baccelli,et al.  On a coverage process ranging from the Boolean model to the Poisson–Voronoi tessellation with applications to wireless communications , 2001, Advances in Applied Probability.

[8]  Konstantinos Kalpakis,et al.  Optimal Placement of Replicas in Trees with Read, Write, and Storage Costs , 2001, IEEE Trans. Parallel Distributed Syst..

[9]  Daniel Kofman,et al.  Analysis and optimization of hierarchical reliable transport protocols , 2001 .

[10]  Pavlin Radoslavov,et al.  Topology-informed Internet replica placement , 2002, Comput. Commun..

[11]  Anthony Claude Busson.  Modélisation spatiale des réseaux , 2002 .

[12]  P. Krishnan,et al.  The cache location problem , 2000, TNET.