Content Placement Algorithms for Cloud-Based Content Delivery Networks

点击翻译按钮获取中文摘要

The advent of elastic resource provisioning in cloud-based networks is largely accepted to be a cost effective solution for content delivery network (CDN) providers. Using the new paradigm of CDN, providers can now lease resources (such as storage, computing and/or bandwidth) in the cloud to build Cloud-based CDNs (CCDNs). Our industrial partner (Ericsson Canada Inc.) is contemplating extending its products and services over its cloud to include Ericsson CCDNs. An intrinsic problem in this kind of networks is the content placement problem. In this project, our team will draw from our experience in algorithm design and in cloud computing to design an efficient push-based content placement model for CCDNs. The design goals of the algorithm are (a) minimizing the cost of leasing storage and bandwidth resources for data coming into and going out of the cloud zones and regions, (b) guaranteeing to the consumers a quality of service defined by a Service Level Agreement (SLA), and (c) minimizing the degree of quality of service violations. The proposed algorithm will incorporate sensitivity to network performance and social interaction metrics. To achieve the objectives of this project, the following milestones will be respected. First, a state-of-the-art report will be provided, using our deep knowledge of the existing content placement algorithms in CDNs. Next, an analytical model design based on optimal linear programming for push-based content placement in CCDNs will be developed and validated using a mathematical solver engine such as CPLEX, lp_solve or GNU-LP Kit. The third step is to design one or multiple heuristic algorithms for push-based content placement in CCDNs and to evaluate their performance by computer-based simulations. Finally, the resulted models and algorithms will be transferred to the industrial partner by technical reports and presented in our meetings. The industrial partner will leverage the results of this project in its existing products and services. This project will establish a highly interesting first research collaboration between the applicant from UQAM and Ericsson Canada Inc. It is highly beneficial for both partners and it is expected that larger scale projects will spin-off.