A Secure, Publisher-Centric Web Caching Infrastructure

The current Web cache infrastructure, though it has a number of performance benefits, does not address many of the publishers’ requirements. We argue that web caches should be enhanced to address publishers’ needs. For example, caches will need to log client accesses, run scripts to dynamically produce content, and give publishers QoS guarantees. In this paper, we propose Gemini, a publishercentric web caching infrastructure. Central to our design is the architectural assumption that the global web cache infrastructure will be heterogeneous, like the Internet itself—caches will belong to many different administrative domains and have different functionalities. The heterogeneous aspect of the infrastructure raises several issues. For example, because caches can alter content, traditional end-to-end security mechanisms can no longer ensure the integrity and authenticity of content. In this paper, we study issues associated with designing such a publish-centric caching infrastructure. In particular, we propose a security architecture that protects publishers and caches from each other in a heterogeneous caching environment. In our design, we ensure that Gemini is incrementally deployable and seamlessly interoperates with the existing caching infrastructure. Along with a system design, we also present experience gained from implementation and preliminary performance results.