Open SkeLLLa opened 8 years ago
it really depends on the transport implementation, and underlying mechanism so plain http is always going to be slower than tcp, for example
@rjrodger I know that http is slow. But it would be great if seneca will have some benchmarks to help developers consider which pros/cons do they have with different kinds of transport.
Yes you're right :)
We need benchmarking script of some kind.
3.4.0 nextgen transport (use option legacy.transport=false) uses keepalive, and is nearly 2X faster.
Hello. I wonder if there are some benchmarks for seneca and it's transport plugins? I've made some tests for myself (with mesh and amqp transports) and they show that it's performance is quiet bad.
PS: whitout any transports it's quiet good. Also results of tcp transport are quet good (on localhost).