I had made some performance ehances for fepss-rpc,you can update.I
had written a performance testcase.
1) run server
http://github.com/bitiboy/fepss-rpc/blob/ab55b7f42e24d6d50098f060679a4273f32e49cb/src/test/java/com/fepss/rpc/RunRpcServerMain.java
2) run test
http://github.com/bitiboy/fepss-rpc
Thank you feedback.Can you share your test code?thanks.
On 5月9日, 上午1时25分, "clint.foster" wrote:
> Many thanks to both authors for providing this code.
>
> I have done some testing with the original protobuf-socket-rpc
> implementation and the MINA-based one. Both work well up to a certain
> numb
On Fri, May 8, 2009 at 9:49 PM, wrote:
> It was necessary to add a length prefix to the request/response
> messages in order to remove the dependency on closing the connection
> to know when the protobuf was finished being read. Also, since
> Message.Builder.mergeFrom() expects to be able to read
I updated the protobuf-socket-rpc implementation so it leaves the
connection open, and I was able to get a 4 or 5 times improvement in
the time it takes to run the test. It also solved the problem of
running out of ephemeral ports.
It was necessary to add a length prefix to the request/response
m
Many thanks to both authors for providing this code.
I have done some testing with the original protobuf-socket-rpc
implementation and the MINA-based one. Both work well up to a certain
number of consecutive calls. However, when I ran a test with 100,000
calls, both implementations started gettin
Added to http://code.google.com/p/protobuf/wiki/RPCImplementations
On Tue, May 5, 2009 at 10:26 PM, Jun Tsai wrote:
>
> Protocol Buffers RPC Server Implemention based on Mina(2.0-M5) and
> Protobuf(2.0.3).
> I had implemented a protocol buffers rpc server.
>
> http://github.com/bitiboy/fepss-rpc