Abstract
Octopus ORAM: An Oblivious RAM with Communication and Server Storage Efficiency
Highlights
As cloud computing has been a common computing paradigm, cloud storage become pervasive
In an Octopus oblivious RAM (ORAM) with de-amortized eviction, as long as s ≥ 8λ and the storage tree has at least 4 leaf nodes, a query process fails with a probability of O(2−λ)
We compare Octopus ORAM to Partition ORAM [37], which is one of the most communication-efficient ORAM that does not require intensive computation or multiple servers and shares the same assumption with Octopus ORAM in that the client has a decent amount of local storage space (in particular,√it assumes the client to have a local storage of O( N ) blocks)
Summary
As cloud computing has been a common computing paradigm, cloud storage become pervasive. Compared to the Path ORAM [38], Octopus ORAM decreases the server storage overhead by 30 times, and has a lower communication cost which is about 23-30% of that of Path ORAM As a tradeoff, it increases the client-side storage, which is affordable to a client that deploys an on-premise facility such as cloud storage gateway. Compared to the S3ORAM [16], the Octopus ORAM with three servers decreases the server storage overhead by 33 times, and halves the client-side communication cost As a tradeoff, it increases the client-side storage, which is affordable to a client deploying an on-premise facility such as cloud storage gateway.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have