Platform for continuous graph update and computation

作者: Aapo Kyrola , Ji Hong , Lidong Zhou , Ming Wu , Xuetian Weng

DOI:

关键词: ComputationGraph databaseDistributed computingSnapshot (computer storage)Stream dataGraph (abstract data type)Computer scienceDatabase transaction

摘要: A continuous stream data (e.g., messages, tweets) is received by ingest nodes of a platform. The may analyze the to create transaction graph updates, assign sequence number transaction, and distribute updates with store from nodes, then report update progresses in progress table. snapshot be taken based on table, graph-mining computation implemented. Tolerate failures decaying supported incremental expansion allowed cope increasing rates needs.

参考文章(54)
Lawrence B. Holder, Maitrayee Mukherjee, Graph-based Data Mining on Social Networks ,(2004)
Bryan B. Tower, Jonathan K. Larson, Database system and related method ,(2008)
Mahadev Konar, Benjamin Reed, Flavio P. Junqueira, Patrick Hunt, ZooKeeper: wait-free coordination for internet-scale systems usenix annual technical conference. pp. 11- 11 ,(2010)
Jinyang Li, Russell Power, Piccolo: building fast, distributed programs with partitioned tables operating systems design and implementation. pp. 293- 306 ,(2010) , 10.5555/1924943.1924964
Chandramohan A. Thekkath, Pradeep Kumar Gunda, Yuan Yu, Lenin Ravindranath, Li Zhuang, Nectar: automatic management of data and computation in datacenters operating systems design and implementation. pp. 75- 88 ,(2010) , 10.5555/1924943.1924949
Anand Rajaraman, Digvijay Singh Lamba, Wang Chee Lam, Taraka Subrahmanya Prasad Siripurapu, Processing data feeds ,(2011)
Eric Newcomer, Philip A. Bernstein, Principles of Transaction Processing ,(1996)