A collaborative transaction coordination model | Un model de coordonare a colaborǎrii tranzacţiilor

In Large-Scale Distributed Collaborative Systems (LSDCS), a large number of various equipment runs cooperatively in a certain relatively vast area. Involved with a large number of transaction activities and processing in a quite different environment, these systems require appropriate transaction management different from its traditional counterparts in many different ways. First, the collaboration system usually consists of multiple distributed autonomous subsystems. Second, transaction tasks are of high concurrency. Third, concurrent transactions' dependence of device resource as well as their different priority reflects the tight collaborations between them. These characteristics make the collaborative resource scheduling in LSDCS a new challenge. To address this issue, we proposed a Collaborative Transaction Coordination Model (CTCM). By introducing the resource reservation idea of Tentative Hold Protocol (THP), we abandoned THP's FIFO algorithm, and used a method of Distributed Constraint Satisfaction Problems (DCSP) modelling transaction environment to optimize the THP's resource booking mechanism. Experimental data indicate that CTCM can significantly improve the efficiency of resource coallocation.