Distributed Locally-Driven Update Commit Algorithm
Original Publication Date: 2003-May-25
Included in the Prior Art Database: 2003-May-25
Related PeopleOther Related People:
A Distributed Replicated System (DRS) consists of a number of computing sites, each providing the same application and containing replicas of the same data items. The problem that typically arises in such a system is the tradeoff between replica consistency and performance (site response time). Replica consistency, or more precisely, inconsistency, is quantified by the misread probability, Pm , and site response time is represented by the waiting time, t, before a read request can be processed. This tradeoff is a general problem in the DRS. The lower the misread probability gets, the longer the waiting time becomes and vice versa. In the ideal case, these two parameters are zero. Since this can never be fulfilled, the general requirement is that both parameters should be minimized as much as possible. A number of techniques have been proposed in the literature coping with this tradeoff.