Goals

The expected innovations of the project include:

Distributed storage platform A secure distributed storage service that crosses administrative and geographical boundaries, and provides data migration/replication and data mobility to adapt to access patterns and network topology with the goal of optimizing latency and bandwidth consumption. The storage is based on a rich object-based model and is adapted to different forms of media. The objects encapsulate metadata for efficient access, search, management and manipulation of data.

Data-intensive computing framework Programming models and a runtime framework enabling secure execution of computational tasks near their data, applicable both to user-driven computational tasks as well as autonomous data derivation and transformation. The high-level programming framework will hide the details of the underlying infrastructure, allow maximal parallelism and generalize on existing frameworks such as map/reduce, Dryad, and Pregel.

Information-centric networks Network architecture and protocols to cache data and computational objects in the network and to implement network-based publish/subscribe mechanisms.

Elastic resource allocation framework A scalable bandwidth-aware platform for scheduling data-intensive computations and services on distributed cloud resources. The platform will allow fine-grain sharing across distributed cloud resources, improve data locality and minimize data movement.

Distributed management framework Global and local resource management services, global resource monitoring services, bandwidth management, a federated identity service, and hybrid integration with public clouds.

Trusted execution environment for protecting a user's services and the provider infrastructure.