The technical part of the research project is planned to be carried out in three phases.
Phase 1: During the first phase, the fellow, in collaboration with the experienced researchers of the Host Organization (HO), will do the following:
- Specify data partitioning algorithms that will yield manageable basic objects
- Specify replication techniques that will enable data availability and fault-tolerance
- Introduce a Distributed Management Service (DMS) that will be responsible for managing data fragmentation and their replication
Phase 2: During this phase the fellow, with the guidance from the scientist in charge (SC), will develop the FCS component of the final DFS. In particular this phase will include the following:
- Embed an efficient atomic R/W object implementation and the specification of the changes that need to be applied to the read and write requests so that they will comply with the architecture introduced in (a) and (b) of Phase 1.
- Perform a theoretical analysis on the impact of the DFS architecture on the operation latency.
Phase 3: During the 3rd phase the fellow will implement the DFS service that results from the combination of the developments in phases 1 and 2, and collect experimental data under the supervision of the experienced researchers of the HO. We expect that ia prototype of the service will be deployed in the following settings:
- Simulation on single processor architecture,
- Multitier real-time planetary-scale environments, and
- High capacity local or wide area environments.