Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Rework the API to use flatbuffers or CapNP depending on which is faster. #11

Open
cholcombe973 opened this issue Sep 1, 2018 · 2 comments

Comments

@cholcombe973
Copy link
Owner

They both seem to satisfy the basic requirements of API versioning, backwards compatibility and fast throughput. With small IO operations it's critical that the API serialize and deserialize quickly so as not to introduce to much latency. Generally 4K writes in Ceph and Gluster are pretty terrible and there's a variety of reasons why but RPC overhead is a big part of it. The other big part is just physics of moving packets to replicas and waiting for the last writer to respond before responding to the client with success.

@cholcombe973 cholcombe973 created this issue from a note in Work tracking (To do) Sep 1, 2018
@cgmcintyr
Copy link

Which of these turned out to be better for use in a DFS?

@cholcombe973
Copy link
Owner Author

I think flatbuffers ended up being the best in terms of serialization speed and network packet size.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
Work tracking
  
To do
Development

No branches or pull requests

2 participants