Implementation and Evaluation ofan

This thesis builds on the work of Fekete et al. ?], who deened an eventually-serializable data service (ESDS) and an abstract algorithm for it. ESDS allows its users to relax consistency requirements in return for improved responsiveness, while providing guarantees of eventual consistency of the replicated data. An important consideration in formulating ESDS was that it could be employed in building real systems. We formulated a framework that assists a programmer in mapping algorithms speciied using I/O Automata notation ?] to distributed implementations. Using the framework, we developed a distributed implementation of ESDS and explored its behavior in a distributed setting. We combined the implementation of ESDS with diierent data types and clients, thus demonstrating the suitability of the service as a general building block. The implementation was experimentally evaluated on a network of workstations. In this setting the implementation scaled in the number of processors and reeected a designed trade-oo between consistency and performance.