Coding and Allocation for Distributed Data Storage: Fundamental Tradeoffs, Coding Schemes and Allocation Patterns

Iryna Andriyanova (1)

1 ETIS - UMR 8051 - Equipes Traitement de l'Information et Systèmes

Distributed storage is one of the most actively developing topics nowadays, thanks to the growing need of storing more and more large amounts of data and to new distributed applications as cloud computing, multimedia content delivery etc. The core of each distributed storage system (DSS) is its erasure-correcting code; the core of the DSS networking protocol is its allocation pattern. With new aforementioned storage applications we are about to attain the limit of well-developed coding schemes. They happen to be inflexible with any size of the storage network and/or very expensive in terms of the storage space or the computational power. The proposed tutorial revisits existing code-allocation solutions for both centralized (i.e. data centers) and decentralized (P2P) storage applications, and gives an insight on how to design more efficient DSSs.

Mots clés: distributed storage; erasure-correcting coding; data allocation

Domaines: Théorie de l'information [cs.IT]

Iryna Andriyanova. Coding and Allocation for Distributed Data Storage: Fundamental Tradeoffs, Coding Schemes and Allocation Patterns. Doctoral. Tutorial at the Swedish Communications Technologies Workshop 2013, Gothenburg, Sweden. 2013. ⟨cel-01263287⟩