julienlau commented on pull request #8544:
URL: https://github.com/apache/pulsar/pull/8544#issuecomment-859446866
> I think this is an important feature to make RabbitMQ/AMQP users that use
this kind of patterns to transition to Pulsar. Would also be nice if it's
integrated into Pulsar functions input.
I switched my PoC of a latency sensitive project from RabbitMQ to Pulsar for
resilience matters.
RabbitMQ is very very fast and efficient at this use case if you are not
using resilient queues (quorum queues).
I took a +10ms by switching to a replicated system (Pulsar) and I think that
if I perform my filtering by adding a second layer of filtered topics I would
experience an additionnal latency hit.
In my view, there could be 2 server side filtering consumer:
- a simple server-side filtering on partition_key would already be a very
very interesting feature.
I mean simple because it should be fast :
- * (star) can substitute for exactly one word.
- # (hash) can substitute for zero or more words.
This would be the *Topic Exchange* of RabbitMQ
https://www.rabbitmq.com/tutorials/amqp-concepts.html
https://www.rabbitmq.com/tutorials/tutorial-five-python.html
- a more advanced filtering on the whole meta-data could also be handy, but
it is generally slower... it may be a good idea to segregate both consumer type
?
It would be like *Headers Exchange* of RabbitMQ
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
[email protected]
