Reliable, efficient, and fast distributed storage systems: theory to practice

Kannan Ramchandran, Rashmi Vinayak, Nihar Shah, Kangwook Lee, Ramtin Pedarsani and Dimitirs Papailiopoulos

Erasure codes such as Reed-Solomon (RS) codes are being extensively deployed in data centers since they offer significantly higher reliability than data replication methods at much lower storage overheads. These traditional codes, however, mandate much higher resources with respect to network bandwidth and disk IO during reconstruction of data that is missing or otherwise unavailable. We research on constructing new storage codes that optimize network and IO resources along with storage, and also build distributed storage systems making use of these new erasure codes.

Figure 1
Figure 1: System model