Location visible to members
This session is on consensus algorithms:
how to best achieve consensus among a set of distributed and unreliable nodes.
- Paxos: family of protocols for reaching consensus among unreliable nodes. Paper 'Paxos made simple' (2001) (http://research.microsoft.com/en-us/um/people/lamport/pubs/paxos-simple.pdf).
- RAFT (1 contrib): protocol explained, scenarios through MSCs, execution of source code: Paper 'In Search of an Understandable Consensus Algorithm' (https://ramcloud.stanford.edu/wiki/download/attachments/11370504/raft.pdf)
- Zab (1 contrib): atomic broadcast layer and protocol used in Zookeeper. Paper 'Zab: High-performance broadcast for primary-backup systems' (http://www.stanford.edu/class/cs347/reading/zab.pdf)
- Viewstamped Replication (1 contrib): replication technique handling node crash failures: Paper 'VR Revisited' (http://pmg.csail.mit.edu/papers/vr-revisited.pdf)
We will also go through latest relevant news on HPC.
- 5 euros to attend
- free access to contributors
- 10 seats available - first arrived first served.
How can I contribute?
- by writing some doc (MD or HTML5)
- by providing Linux install scripts (Bash, Python, ..)
- by providing test source code (benchmarks, etc..)
Contribution needed on 'Paxos'.
Can you contribute on this?