Abstract | ||
---|---|---|
This paper reports our experiments to compare various deployment strategies of memcached-like in-memory storage for Hadoop on supercomputers, where each node often does not have a local disk but shares a slow central disk. For the experiments, we developed our own memcached-like file system, named SEMem, for Hadoop. Since SEMem was designed for supercomputers, it uses MPI for communication. SEMem is configurable to adopt various deployment strategies and our experiments revealed that a good deployment strategy was allocating some nodes that work only for in-memory storage but do not directly perform map-reduce computation. |
Year | Venue | Field |
---|---|---|
2017 | Euro-Par | File system,Software deployment,Computer science,Distributed computing,Computation |
DocType | Citations | PageRank |
Conference | 0 | 0.34 |
References | Authors | |
8 | 2 |
Name | Order | Citations | PageRank |
---|---|---|---|
Thanh-Chung Dao | 1 | 0 | 0.34 |
Shigeru Chiba | 2 | 1281 | 140.78 |