Mailing-List: contact geronimo-dev-help@incubator.apache.org; run by ezmlm
Precedence: bulk
Reply-To: geronimo-dev@incubator.apache.org
Date: Fri, 8 Aug 2003 15:32:07 -0400 (EDT)
From: Aaron Mulder <ammulder@alumni.princeton.edu>
To: geronimo-dev@incubator.apache.org
Subject: Re: RES: RES: "Virtual Hosting"
In-Reply-To: <5.1.0.14.2.20030808125623.00a80830@pop.ncsa.uiuc.edu>
Message-ID: <Pine.LNX.4.44.0308081528590.5667-100000@www.princetongames.org>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII

On Fri, 8 Aug 2003, Michael Remijan wrote:
> I thought it would be assumed that the NIO packages would be the way to go.

	Well, that's really a different question.  Question 1 is, should 
we do something like I describe below?  Question 2 is, regardless of the 
answer to question 1, should we write a bunch of NIO code to implement 
our socket handling?

	I heard the initial releases of NIO were fairly buggy, but I've 
also heard that 1.4.2 made dramatic strides in addressing that.  I don't 
have any personal experience either way.  I'd certainly be open to going 
that way.  But if we don't want to do what's described below, we can start 
by just relying on Sun's RMI server implementation or something along 
those lines.

Aaron

> At 03:18 PM 8/8/2003 -0400, you wrote:
> 
> >         Which reminds me, does anyone have any thoughts on the possibility
> >and/or wisdom of multiplexing all network communication over a single
> >port?  You know, HTTP, RMI, JNDI, IIOP, etc. all on the same port (sure, I
> >have 7001 in mind)?  It makes it darn easy to configure, firewall, tunnel,
> >etc. at the cost of (I assume) quite a bit of complexity in terms of
> >initially accepting the connection, figuring out the protocol, and sending
> >it to the proper handler.
> 
>