Computer Science Colloquia
Thursday, May 16, 2013
Advisor: Sudhanva Gurumurthi
Attending Faculty: Mary Lou Soffa (Chair) and Mircea R. Stan
3:00 PM, Rice Hall, Rm. 242
Master's Thesis Presentation
Accelerating Reliability Simulation of NAND-Flash Based Solid State Drives
Reliability is an important factor to consider when designing and deploying SSDs in storage systems. Both the endurance and the retention time of flash memory are impacted by the history of low-level stress and recovery patterns to Flash cells, which are determined by the workload characteristics, the time period over which the workload utilizes the SSD, and the FTL algorithms. Accurately assessing SSD reliability requires simulating the workload behavior over timescales that spans several years, which is very time-consuming. In this thesis, we present a methodology that uses sampling and clustering techniques to reduce the simulation time while maintaining high accuracy. The methodology relies on the key insight that most of the large changes in retention time occur early in the lifetime of the SSD, whereas most of the simulation time is spent in the latter stages. This allows for simulation acceleration to be focused on the latter stages without significant loss of accuracy. We show that our approach provides an average speed up of 12X over detailed simulation with an error of 3.21% in the estimated mean and 6.42% in the estimated standard deviation of the retention times of the blocks in the SSD.