Cloudifying Applications for Community Networks

The objective and work here will be to evaluate a baseline system/server and then propose, prototype and evaluate the "cloudification" of typical community services and as part of a community cloud (with is dynamics and other characteristics). Ideas about possible "cloudification" projects:

• A latency sensitive service: e.g. taking a mature Open Source VOIP or video streaming service and exploring its "cloudification" so that it could adapt to changing environmental conditions (resources that degrade and can be congested, varying net/node latency), users that move around the network or changes in the number of users (overall load). That implies the elastic organisation of an overlay of VOIP or streaming servers.
• Another examples would be evaluating systems such as Hadoop or another typical cluster system (e.g. a data/event streaming service, a replicated and highly available file system such as Lustre.
• A latency insensitive service: an elastic cloud-based email service that can be elastic in terms resource claims with respect to the amount and load of spam to check, flow of legitimate mails, delivery time, etc.

Some context information about Community Lab and Community Clouds.

Supervised by: Leandro Navarro <leandro@ac.upc.edu>