Task Management And Protein Folding On The Distributed Computing Platform Of P2HP

In the area of bioinformatics, people have widely accepted the natural idea, getting a structure of protein by stimulation computation, which requires a huge mount of computation ability. To solve the lack of computing ability, many distributed platforms have been proposed and constructed for the computation of protein folding. Although these platforms have resolved the lack of computation resources at some degrees, they always have two shortcomings: a hidden trouble the server will break down, the task partition is not flexible. According the problems listed above, the design and implementation of P2HP (P2P-based high performance computing platform) conquers these problems, making the platform more flexible, fault-tolerant and scalable.To P2HP, high efficient mechanism of task management is indispensable. To reach this aim, P2HP organizes the task by the type of files, which decide there are some merits in the task storing, the task transmission and the task operation. P2HP takes the strategy of storing the metadata and the embedded database to improve the efficiency of task access, defines the transmission protocol based on the data pipe to quick the task transmission, provides a integrated development software package to convenience the operations for users and volunteers.By analyzing the algorithm of PERM (Pruned-Enriched-Rosenbluth Method) based on the lattice model, we design a strategy to distribute this algorithm and give some ideas about how to implement the strategy on some distributed platforms. At last, the PERM algorithm is deployed on P2HP. The success of running the PERM on the platform will incarnate the generalization, usability and efficiency of P2HP and provide a new way to solve the problem of protein folding.The experiment result demonstrates the distributed operation of protein folding computing based on PERM by using P2HP, not only can increase the amount of searching the result space, but also be able to find the best structure of protein molecular in a shorter time, because the speedup of the P2HP will be enhanced when more volunteers join in the platform.