Abstract:TransCom is a new distributed system which allows users to load programs remotely on demand. The cache plays an important role in the TransCom system design but existing cache simulation methods can not provide accurate TransCom cache performance analysis. Therefore, a shared storage cache simulator (SSCS) algorithm was developed which extends the traditional stack-distance model by introducing user management. The system identifies each cache block by a vector composed by the block id and user id and then separately manages, schedules and measures the block performance according to the user behavior. Cache performance tests verify the simulations using the algorithm are accurates. Optimization based on the conclusions greatly improves system performance and scalability.

Simulation Analysis and Validation of Cache Performance in TransCom Systems