Counterintuitive Characteristics of Optimal Distributed LRU Caching Over Unreliable Channels

Joint optimization of cache placement and request routing in unreliable networks

“…However, the main limitation is such approaches overlook the bandwidth requirements on the delivery paths, which hinder a wider applicability within bandwidth-scarce environments. Additionally, [25] aims to cache replications of the same data in multiple caches to optimize the LRU caching in an unreliable network environment. However, it only considers single-hop routing in its distributed caching system, and thus cannot be directly applied into the path routing with multi-hops.…”

“…Providers thus face pressure to provision uninterrupted reliable services while reducing operational costs due to significant software and hardware failures [22,23,24]. The distributed cache networks where many small servers deployed at edge nodes are more prone to failure caused by power supply shutdowns, hardware failures [25,26]. Due to further indexing and transmission distance, network failure events may lead to the failure of cooperative caching to effectively support delay-sensitive services [27,28].…”