Low latency replication coded storage overmemory-constrained servers

# 46


We consider a distributed storage system storing asingle file, where the file is divided into equal sized fragments. The fragments are replicated with a common replication factor and stored across servers with identical storage capacity. An incoming download request for thisfile is sent to all the servers, and it is considered serviced when all theunique fragments are downloaded. The download time for all fragments across allservers, is modeled as an independent and identically distributed (i.i.d.)random variable. The mean download time can be bounded in terms of the expectednumber of useful servers available after gathering each fragment. We find themean number of useful servers after collecting each fragment, for a random storage scheme for replication codes. We show that the performance of therandom storage for replication code achieves the upper bound for expectednumber of useful servers at every download asymptotically in number of servers for any storage capacity and performance of this storage scheme to that of Maximum Distance Separable (MDS) coded storage. This is a joint work with AjayBadita, Pradeep Sarvepalli and Parimal Parag.

Rooji Jinan, IISc Bangalore

Rooji Jinan is a 4th year PhD student at Robert Bosch Centre forCyber Physical Systems, IISc. She is guided by Dr. Parimal Parag, Department ofElectrical Communication Engineering, IISc. Her current research focusses inareas of low latency communication for cyberphysical systems and low latency distributed storage and computation. Rooji received her B.Tech. degree in Electronics and Communication and M.Tech. in Communication Engineering and Signal Processing from CalicutUniversity, Kerala.