Implementation of director-aggregator gRPC communication #306
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
aleksandr-mokrov
requested review from
dmitryagapov,
igor-davidyuk and
psfoley
dmitryagapov
suggested changes
Feb 2, 2022
psfoley
reviewed
Feb 2, 2022
dmitryagapov
approved these changes
Feb 4, 2022
tests/github/interactive_api_director/experiments/pytorch_kvasir_unet/envoy/envoy_config.yaml
Outdated
Show resolved
Hide resolved
aleksandr-mokrov
force-pushed
the
director-aggregator-communication-by-rpc
branch
from
5d66d6e to
954d675
Compare
Signed-off-by: Aleksandr Mokrov <[email protected]> Update openfl/federated/plan/plan.py Co-authored-by: Igor Davidyuk <[email protected]>
aleksandr-mokrov
force-pushed
the
director-aggregator-communication-by-rpc
branch
from
33a14fb to
d5ec206
Compare
|
@aleksandr-mokrov can you please resolve the conflicts and request for review again? |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This is implementation of Director - Aggregator communication by gRPC in order to provide possibility of a separate launch.
1 . There was implemented the next APIs: GetMetricStream, GetTrainedModel, GetExperimentDescription.
2. All request from user sent to the Director service. The Director service checks a caller, an experiment status and etc, and if it possible sends request to Aggregator service. User -> Federation -> director client -> Director service -> Director -> aggregator client -> Aggregator service -> Aggregator.
3. It was implemented an asynchronous aggregator client (AsyncAggregatorGRPCClient) for asynchronous Director based on the synchronous client.
4. The async aggregator client makes several attempts to connect to the server in case of unavailability (each second during 30 seconds by default; synchronous aggregator client, that used by collaborators makes it last forever). It implemented using Interceptors.
5. Aggregator's IP address and ports are taken from the plan (a port is generated by using the plan hash).
6. pytorch_kvasir_unet test was fixed.
There is a restriction: it is possible get an information from an aggregator only when it is alive. It is acceptable for GetMetricStream, but should be resolved for GetTrainedModel in the future (there several way to resolve, but they should be discussed).