An Age-Threshold Algorithm for Garbage Collection in Log-Structured Arrays and File Systems

In this paper, we propose and study a new algorithm for choosing segments for garbage collection in Log-Structured File Systems (LFS) and Log-Structured Arrays (LSA). We compare the performance of our new algorithm against previously known algorithms such as greedy and cost-beneet through simulation. The basic idea of our algorithm is that segments which have been recently lled by writes from the system should be forced to wait for a certain amount of time (the age-threshold) before they are allowed to become candidates for garbage collection. The expectation is that if the age-threshold is properly chosen, segments that have reached the age-threshold are unlikely to get signiicantly emptier due to future rewrites. Among segments that pass the age-threshold and become candidates for garbage collection, we select ones that will yield the most amount of free space. We show, through simulation, that our age-threshold algorithm is more eecient at garbage collection (produces more free space per garbage-collected segment) than greedy or cost-beneet; this means that designs using age-threshold will give better system performance than designs using greedy or cost-beneet. It is also simpler to implement a scalable version of the age-threshold algorithm than to implement a scalable version of the cost-beneet algorithm. The performance of the age-threshold algorithm depends on good choice of an age-threshold; therefore, we also give an analysis which can be used to choose an optimal age-threshold under certain workload assumptions. We also suggest how to choose good age-thresholds when nothing is known about the workload.