Slicedup: a tenant-aware memory deduplication for cloud computing
Fernando Vañó-García, Héctor Marco-Gisbert · The UWS Academic Portal (University of the West of Scotland) · 2018
Memory deduplication allows cloud infrastructure providers to increase the profit of memory resources by taking advantage of the redundant nature of virtual machines footprint. Although it is an important feature to manage the memory resources of a cloud system efficiently, unfortunately, it enables different types of side-channel attacks which, in practice, means disabling memory deduplication. In this paper, we present Slicedup, a tenant-aware memory deduplication mechanism that prevents side-channel attacks. Our proposal enables cloud providers to get the deduplication saving benefits while preventing side-channel attacks among tenants. Since Slicedup is a design-solution, it can be implemented in any operating system, regardless of its version, architecture or any other system dependence. Finally, we show how Slicedup prevents side-channel attacks while providing similar memory savings when the number of tenants per physical host is low.