Optimal Caching for Low Latency in Distributed Coded Storage Systems

Erasure codes have been widely considered as a promising solution to enhance data reliability at low storage costs. However, in modern geo-distributed systems, erasure may incur high access latency they require retrieval from multiple remote nodes. This hinders the extensive application of data-intensive applications. paper proposes novel caching schemes achieve distributed coded systems. Assuming that future popularity and network information are available, an offline scheme is proposed explore optimal for latency. The categorizes all feasible decisions into set cache partitions, then obtains decision through market clearing price each partition. Furthermore, guided by scheme, online according measured real time, without need completely override existing decisions. Both theoretical analysis experiment results demonstrate can approximate well with dramatically reduced computation complexity.