You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
A way to aggregate multiple messages into one/fewer may be a nice option for apps operating on CRDTs or other data types which are nice to 'merge'. It could enable massive savings in terms of network bandwidth and cpu processing for certain use cases.
One particular use case I have in mind is voice-over-pubsub app, where an Aggregator would mix multiple sound channels into one (sort of audio-crdt), reducing network/cpu load for participants of big group calls.
I'm not sure about implementation details, one problem is probably a trade-off between efficiency and latency (e.g for voice apps we ideally don't want to increase latency more than 1-whatever-length-audio-chunk), so the optimal interface would let the aggregator control that well.
Aggregator isn't the best name, and the interface optimally wouldn't require it to run in a separate goroutine to use. It would probably plug next to Validator stuff (and could be used to implement it)
The text was updated successfully, but these errors were encountered:
A way to aggregate multiple messages into one/fewer may be a nice option for apps operating on CRDTs or other data types which are nice to 'merge'. It could enable massive savings in terms of network bandwidth and cpu processing for certain use cases.
One particular use case I have in mind is voice-over-pubsub app, where an Aggregator would mix multiple sound channels into one (sort of audio-crdt), reducing network/cpu load for participants of big group calls.
I'm not sure about implementation details, one problem is probably a trade-off between efficiency and latency (e.g for voice apps we ideally don't want to increase latency more than 1-whatever-length-audio-chunk), so the optimal interface would let the aggregator control that well.
Aggregator
isn't the best name, and the interface optimally wouldn't require it to run in a separate goroutine to use. It would probably plug next to Validator stuff (and could be used to implement it)The text was updated successfully, but these errors were encountered: