Title | ||
---|---|---|
A Novel Object Placement Protocol For Minimizing The Average Response Time Of Get Operations In Distributed Key-Value Stores |
Abstract | ||
---|---|---|
We present a novel object placement protocol for distributed storage systems that migrates objects between nodes in run time with the goal to minimize the average response times in the system. We rely on a combination of consistent hashing with small lookup tables for objects that have been moved and for which the hash function cannot be aware of. We test our approach in various scenarios based on the assumptions that "get" operations follow a power law distribution and that the request rate is the most significant contributor in the decreasing of the response time based on past research. The results show significant improvements in comparison to the baseline scenario. |
Year | DOI | Venue |
---|---|---|
2017 | 10.1109/BigData.2017.8258300 | 2017 IEEE INTERNATIONAL CONFERENCE ON BIG DATA (BIG DATA) |
Keywords | DocType | ISSN |
distributed storage systems, consistent hashing, directory look-ups, power (Zipf) law distribution | Conference | 2639-1589 |
Citations | PageRank | References |
0 | 0.34 | 0 |
Authors | ||
3 |
Name | Order | Citations | PageRank |
---|---|---|---|
Antonios Makris | 1 | 8 | 2.56 |
Konstantinos Tserpes | 2 | 276 | 38.63 |
Dimosthenis Anagnostopoulos | 3 | 222 | 42.72 |