Items from Performance, Scalability and Availability track

Anycast-based content delivery networks (CDNs) have many properties that make them ideal for the large scale distribution of content on the Internet. However, because routing changes can result in a change of the endpoint that terminates the TCP session, TCP session disruption remains a concern for anycast CDNs, especially for large file downloads. In this paper we demonstrate that this problem does not require any complex solutions. In particular, we present the design of a simple, yet efficient, mechanism to handle session disruptions due to endpoint changes. With our mechanism, a client can continue the download of the content from the point at which it was before the endpoint change. Furthermore, CDN servers purge the TCP connection state quickly to handle frequent switching with low system overhead. We demonstrate experimentally the effectiveness of our proposed mechanism and show that more complex mechanisms are not required. Specifically, we find that our mechanism maintains high download throughput even with a reasonably high rate of endpoint switching, which is attractive for load balancing scenarios. Moreover, our results show that edge servers can purge TCP connection state after a single timeout-triggered retransmission without any tangible impact on ongoing connections. Besides improving server performance, this behavior improves the resiliency of the CDN to certain denial of service attacks.

Web hosting providers are increasingly looking into dynamic hosting to reduce costs and improve the performance of their platforms. Instead of provisioning fixed resources to each customer, dynamic hosting maintains a variable number of application instances to satisfy current demand. While existing research in this area has mostly focused on the algorithms that decide on the number and location of application instances, we address the problem of efficient enactment of these decisions once they are made. We propose a new approach to application placement and experimentally show that it dramatically reduces the cost of application placement, which in turn improves the end-to-end agility of the hosting platform in reacting to demand changes.

This paper proposes and evaluates a Network Aware Forward Caching approach for determining the optimal deployment strategy of forward caches to a network. A key advantage of this approach is that we can reduce the network costs associated with forward caching to maximize the benefit obtained from their deployment. We show in our simulation that a 37% increase to net benefits could be achieved over the standard method of full cache deployment to cache all POPs traffic. In addition, we show that this maximal point occurs when only 68% of the total traffic is cached. Another contribution of this paper is the analysis we use to motivate and evaluate this problem. We characterize the Internet traffic of 100K subscribers of a US residential broadband provider. We use both layer 4 and layer 7 analysis to investigate the traffic volumes of the flows as well as study the general characteristics of the applications used. We show that HTTP is a dominant protocol and account for 68% of the total downstream traffic and that 34% of that traffic is multimedia. In addition, we show that multimedia content using HTTP exhibits a 83% annualized growth rate and other HTTP traffic has a 53% growth rate versus the 26% over all annual growth rate of broadband traffic. This shows that HTTP traffic will become ever more dominent and increase the potential caching opportunities. Furthermore, we characterize the core backbone traffic of this broadband provider to measure the distance travelled by content and traffic. We find that CDN traffic is much more efficient than P2P content and that there is large skew in the Air Miles between POP in a typical network. Our findings show that there are many opportunties in broadband provider networks to optimize how traffic is delivered and cached.