Hi Brian,

There is no underlying statically defined pool.
If you are using the same http client instance for 2 separate servers, you
can define independently the number or connections to keep in the pool for
each server as explained here:
http://hc.apache.org/httpcomponents-client-ga/tutorial/html/connmgmt.html#d5e627
So there is no need to use 2 seperate http client instances.


2013/6/19 Brian <[email protected]>

> Hi All,
>
> I have a question about Apache Http Client use in J2EE application.
>
> I am working on a legacy app which connects to two different http servers
> and it currently uses Multi Threaded Connection Manager backed, singleton
> instance of Http Client declared statically to manage http requests to both
> http servers. The transactional load between the two varies significantly,
> so we are noticing that multithreaded http connection manager gets into a
> waiting state under load, as all connections are in use.
>
> My question - is it a good design decision to create separate Http Client
> object instance for each underlying http server, so that connection pool
> sizes can be tuned independently of one another and transactions impacts
> isolated, so that if the load towards one server increases the transactions
> towards the other are not negatively impacted from http connection point of
> view? I have read the docs on multi threading, and in the docs there is a
> recommendation on using a single Http Client per application, thus my
> question.
>
> I also just want to ensure that the API does not have statically defined
> common underlying connection pool management code, which might cause
> performance bottlenecks if multiple instances of Http Client are used.
>
>
>
> Thanks
>
>
>
> Brian
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [email protected]
> For additional commands, e-mail: [email protected]
>
>

Reply via email to