Liberating Content: Adaptive Middleware for Providing Content-Centric Networking Support

This paper investigates the role of middleware in deploying contentcentric networks. A middleware called Juno is described, exploiting recent observations that applications rarely have an interest in how or where their content (e.g. videos, software updates) is obtained from as long as it is delivered within certain requirement bounds. Juno provides a high level abstraction through which applications can request content agnostic to its underlying delivery. This abstraction subsequently allows dynamic (re)configuration, through software reflection, based on flexible and extensible runtime meta-data. The entire process involves the location of content sources alongside the selection of the most optimal available delivery scheme. The middleware then reconfigures to interact with this choice, providing transparent interoperation for the application. Juno has been evaluated by deploying a number of casestudies on the Emulab testbed. It is shown that traditional applications, deployed with static, design-time selected delivery schemes cannot compete with Juno’s approach of dynamic selection. Further, through distributed adaptation it is shown how Juno can adapt to environmental variations.