Vikrant, how much latency is between you and the server? I doubt fork join would help at all, I'm sure the latency is successive server round trips. I thought TreeCache was already pretty decent on that front, because we use the async version of every call. I assumed doing that would allow me to "pipeline" server requests and react as each response came back. But it occurs to me that the connection might actually be blocking per request under the hood. How does that stuff work there?
On Wed, Sep 16, 2015 at 12:53 PM, Jordan Zimmerman < [email protected]> wrote: > Another idea would be to modify TreeCache to use fork/join. I may be > overthinking things here. > > -JZ > > > > On September 16, 2015 at 11:52:07 AM, Vikrant Singh ( > [email protected]) wrote: > > Tree is not quite big.. it is only 4 level.. .and nodes on 3rd level will > have only 1 child node. > I am interested only in monitoring 3 & 4 level.. 1 & 2 will never change. > > Even after changing the MaxDepth (from default to 3) i do not see much > performance gain. > > Any advise to get some performance here? > > On Tue, Sep 15, 2015 at 7:37 PM, Jordan Zimmerman < > [email protected]> wrote: > >> I don't think you'll be able to extend the existing TreeCache to handle >> multi transactions. It would need to be a modification within Curator. >> >> >> Hmm - that might be a nice feature. I’ll think about it. >> > >
