Graph An

From Master Projects
Revision as of 13:51, 25 November 2014 by Paboncz (talk | contribs) (New page: {{Projectproposal |Contact person=Peter Boncz |Master areas=High Performance Distributed Computing, Parallel and Distributed Computer Systems |Project |Fulfilled=No ...)

(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to: navigation, search

About Graph An


Graph Analysis on large datasets is a hot topic in Big data processing.

Graph Programming frameworks like Giraph and Graphlab and associated languages such as Green Marl currently heither have non-scalable or relatively slow back-ends. This project aims to see if high-performance database query processing techniques could be used to power a graph programming framework. A frst question would be to code and test a number of graph algorithms (PageRank,APSP,Clustering) in terms of iterative dtabase queries. This chould bpreferably be developed and tested on scalable parallel database systems, such as Impala or Vectorwise-on-Hadoop. A second question would be whether a language like Green Marl could be automatically translated to such query plans, i.e. the task would be to create a graph query compiler.