Re: SigleTon definitive question
From: Ken Kolda (ken.kolda_at_elliemae-nospamplease.com)
Date: 11/10/04
- Next message: Joannes Vermorel: "Making the framework use a custom IFormatter for Remoting"
- Previous message: Ken Kolda: "Re: Maintaining CAO in Web Service"
- In reply to: Pedro Gago: "SigleTon definitive question"
- Next in thread: Pedro Gago: "Re: SigleTon definitive question"
- Reply: Pedro Gago: "Re: SigleTon definitive question"
- Messages sorted by: [ date ] [ thread ]
Date: Wed, 10 Nov 2004 09:32:09 -0800
Based on your persistence, I think you're looking for someone to tell you
exactly where the time's going, but I don't think anyone can give you
anything more than a few educated guesses. For example, somone already
mentioned it could be in JITing some classes that get used for the first
time, it could be in loading/initializing the TCP/IP stack, loading some
external DLLs, in DNS resolution or a number of other places. If you were
saying that you were losing a second or two on the first call, I'd
understand your desire to know the source, but you've said it was only about
a tenth of a second. I think you're going to be wasting your time trying to
find the exact location where this time is spent.
By the way, why do you unregister your channel after each call? Is this or
the sake of testing the performance or do you have a requirement that this
connection be as short-lived as possible?
Ken
"Pedro Gago" <pgago_y@yahoo.es> wrote in message
news:%23rknOgwxEHA.1292@TK2MSFTNGP10.phx.gbl...
> Hi everybody,
>
> I'm doing tests with singleton activation mode and I have a doubt that I
> can't solve.
>
> With two clients, the first function call to a server component wastes
more
> or less the same time in both clients. Someone said to me in this news
group
> that this time doesn't depend on the server object creation (because both
> use the same instance) but on the time to open the network connection with
> the server.
>
> But, if I have only one client in which I open and close the channel
every
> time, I can see that the time of making again the connection is 0
> miliseconds.
>
> So, in all this approach there is something wrong. Can anybody help me?
>
> Thanks.
>
> Pedro Gago
>
>
- Next message: Joannes Vermorel: "Making the framework use a custom IFormatter for Remoting"
- Previous message: Ken Kolda: "Re: Maintaining CAO in Web Service"
- In reply to: Pedro Gago: "SigleTon definitive question"
- Next in thread: Pedro Gago: "Re: SigleTon definitive question"
- Reply: Pedro Gago: "Re: SigleTon definitive question"
- Messages sorted by: [ date ] [ thread ]
Relevant Pages
|