Lossy distributed storage with limited failure loss
Minh Thành Vu, Tobias J. Oechtering, Mikael Skoglund · 2015
In this paper a successive refinement problem is considered where decoders can access only distributed storage nodes, subject to an additional constraint on the loss of information if failures happen. We characterize the optimal trade-off for a given a distortion tuple and an equivocation constraint between storage nodes. A binary example is provided to illustrate the trade-off between the loss of information constraint and the corresponding achievable rate region.