Along with the spread of the World Wide Web, social networks and the Internet of Things in the past decades, the need for systems and algorithms that can process massive graphs has been continuously increasing. There has been considerable amount of research done in distributed graph processing since the emergence of such large-scale graphs. Another steadily growing field in the past years has been stream processing. This rise of interest can be attributed to the need to process large amounts of continuously streaming data with scalability, fault tolerance and very low latency. Graph streaming, the unification of these two fields is a rather new idea, with some research already being done on it. Processing graphs that are unbounded, and so large that they cannot be stored in memory or even on the disk, is only possible with a distributed graph streaming model. Our goal is to provide a graph streaming model and API that can handle common transformations and provide statistics on streamed graphs. This graph streaming API is created on top of Flink streaming and provides similar interfaces to Gelly, which is the graph library on the batch processing part of Flink.
| Identifer | oai:union.ndltd.org:UPSALLA1/oai:DiVA.org:kth-170425 |
| Date | January 2015 |
| Creators | Bali, János Dániel |
| Publisher | KTH, Skolan för informations- och kommunikationsteknik (ICT) |
| Source Sets | DiVA Archive at Upsalla University |
| Language | English |
| Detected Language | English |
| Type | Student thesis, info:eu-repo/semantics/bachelorThesis, text |
| Format | application/pdf |
| Rights | info:eu-repo/semantics/openAccess |
| Relation | TRITA-ICT-EX ; 2015:169 |
Page generated in 0.0915 seconds