Content Delivery Networks (CDNs) such as Akamai and Mirror Image place web server clusters in numerous geographical locations to improve the responsiveness and locality of the content it hosts for end-users. However, their services are priced out of reach for all but the largest enterprise customers. An alternative approach to content delivery could be achieved by leveraging existing infrastructure provided by ‘Storage Cloud ’ providers, who offer internet accessible data storage and delivery at a fraction of the cost. In this paper, we introduce MetaCDN, a system that exploits ‘Storage Cloud ’ resources, creating an integrated overlay network that provides a low cost, high performance CDN for content creators. MetaCDN removes the complexity of dealing with multiple storage providers, by intelligently matching and placing users ’ content onto one or many storage providers based on their quality of service, coverage and budget preferences. MetaCDN makes it trivial for content creators and consumers to harness the performance and coverage of numerous ‘Storage Clouds ’ by providing a single unified namespace that makes it easy to integrate into origin websites, and is transparent for end-users. We then demonstrate the utility of this new approach to content delivery by showing that the participating ‘Storage Clouds ’ used by MetaCDN provide high performance (in terms of throughput and response time) and reliable content delivery for content consumers. 1

Abstract. Content Delivery Networks (CDNs) are commonly believed to offer their customers protection against application-level denial of service (DoS) attacks. Indeed, a typical CDN with its vast resources can absorb these attacks without noticeable effect. This paper uncovers a vulnerability which not only allows an attacker to penetrate CDN’s protection, but to actually use a content delivery network to amplify the attack against a customer Web site. We show that leading commercial CDNs – Akamai and Limelight – and an influential research CDN – Coral – can be recruited for this attack. By mounting an attack against our own Web site, we demonstrate an order of magnitude attack amplification though leveraging the Coral CDN. We present measures that both content providers and CDNs can take to defend against our attack. We believe it is important that CDN operators and their customers be aware of this attack so that they could protect themselves accordingly. 1

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.

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.

Content delivery networks (CDNs) have become a crucial part of the modern Web infrastructure. This paper studies the performance of the leading content delivery provider – Akamai. It measures the performance of the current Akamai platform and considers a key architectural question faced by both CDN designers and their prospective customers: whether the co-location approach to CDN platforms adopted by Akamai, which tries to deploy servers in numerous Internet locations, brings inherent performance benefits over a more consolidated data center approach pursued by other influential CDNs such as Limelight. We believe the methodology we developed for this study will be useful for other researchers in the CDN arena.

This paper proposes a new delivery-centric abstraction, which extends the existing content-centric interface. Specifically, a delivery-centric interface allows applications to generate content requests agnostic to location or protocol, with the additional ability to stipulate high-level requirements (e.g. performance, encryption support). Fulfilling these requirements, however, is complex as often the ability of a provider to satisfy requirements will vary between different consumers and over time. To address this, we also present a middleware, Juno, which is capable of re-configuring between the use of different content protocols and infrastructure. Juno exploits this to make request-time decisions on how best to access content, thus liberating developers from making design-time decisions.

Abstract—Cloud-computing systems are rapidly gaining momentum, providing flexible alternatives to many services. We study the Domain Name System (DNS) service, used to convert host names to IP addresses, which has historically been provided by a client’s Internet Service Provider (ISP). With the advent of cloud-based DNS providers such as Google and OpenDNS, clients are increasingly using these DNS systems for URL and other name resolution. Performance degradation with cloud-based DNS has been reported, especially when accessing content hosted on highly distributed CDNs like Akamai. In this work, we investigate this problem in depth using Akamai as the content provider and Google DNS as the cloud-based DNS system. We demonstrate that the problem is rooted in the disparity between the number and location of servers of the two providers, and develop a new technique for geolocating data centers of cloud providers. Additionally, we explore the design space of methods for cloudbased DNS systems to be effective. Client-side, cloud-side, and hybrid approaches are presented and compared, with the goal of achieving the best client-perceived performance. Our work yields valuable insight into Akamai’s DNS system, revealing previously unknown features. I.