RAFT at work: speeding-up mapreduce applications under task and node failures.
Jorge-Arnulfo Quiané-RuizChristoph PinkelJörg SchadJens DittrichPublished in: SIGMOD Conference (2011)
Keyphrases
- node failures
- fault tolerance
- data replication
- distributed computing
- congestion control
- overlay multicast
- overlay network
- fault tolerant
- cloud computing
- distributed systems
- end to end
- distributed databases
- load balancing
- peer to peer
- window size
- response time
- packet loss
- high availability
- network bandwidth
- data management
- end to end delay
- real time