Research Training Group SCARE

Awais Usman: A Novel Highly Available Data
Replication Strategy exploiting Data Semantics, Coding Techniques and
Prior At-Hand Knowledge

Data replication is an important research area, as reliable access to data makes up the base of most of IT services. High operation availabilities, low operation costs and data consistency are major target conflicts in almost every data replication research. In this paper, we introduce a data semantics and data encoding-based data replication strategy called Semantic Data Replication (SDR). SDR focuses on the exploitation of data semantics, coding schemes, and efficient search algorithms to come up with a strategy whose goal is to provide high operation availabilities with low costs. SDR primarily takes advantage of the finite state space introduced by the coding scheme for the replicated objects, and it uses a priori knowledge from the finite state space to know in advance of the state transformation, which makes this strategy more efficient than existing infinite state space replication strategies. Exploitation of the data coding schemes is a significant contribution in SDR, as the coding scheme helps in fast execution of read and write operations. It provides consistent access to the state of the replicated object for the read operations.