|
@jepsen_io | |||||
|
On Wednesday, August 14, I'll be giving a livestreamed talk with @fauna, discussing various approaches for transactional isolation in globally distributed systems, including Spanner, @CockroachDB, FaunaDB, @YugaByteDB, @VoltDB, Percolator, and TiDB.
www2.fauna.com/wcdistributeda…
|
||||||
|
||||||
|
Jepsen
@jepsen_io
|
15. kol |
|
And here's the video! youtube.com/watch?v=w_zYYF…
|
||
|
|
||
|
Nathan VanBenschoten
@natevanben
|
18. stu |
|
Nice video, thanks again Kyle! Upon a rewatch, I'm left wondering whether I'm missing something on the Calvin Lamport diagram: youtu.be/w_zYYF3-iSo?t=….
|
||
|
|
||
|
Nathan VanBenschoten
@natevanben
|
18. stu |
|
Doesn't "Wait for all log shards" imply a third sync cross-DC hop? For any log segment without a local Paxos leader, won't a given DC need to wait for propose, ack, AND commit to be able to append its txns to the global txn log, which is a prereq to apply any later txn? cc @evan
|
||
|
|
||
|
Jepsen
@jepsen_io
|
31. srp |
|
I don't think Dgraph advertises itself for use in globally distributed contexts--AFAIK it's only intended for single-datacenter deployments. I might be wrong, though!
|
||
|
|
||
|
Ryan Worl
@ryanworl
|
12. kol |
|
Why isn’t FoundationDB included in this list? If it is because you think it works like Percolator, I’d be happy to explain all the ways it is different. The architecture is extremely interesting. Let me know if you need any information. cc @evan @dhruv_gupta
|
||
|
|
||
|
Jepsen
@jepsen_io
|
12. kol |
|
I would love to, but I had limited time to research. Maybe I can do another version of this talk with FDB later!
|
||
|
|
||