[Ovirt-devel] [R&D] Breaking the Browser

[Jason Guiditta]

On Thu, 2008-07-03 at 07:22 -0400, Bryan Kearney wrote:
> 
> Jason Guiditta wrote:
> > ==Part 1: The Problem ==
> > 
> > I have been asked to look into some options for how we might fulfill a 
> > requirement for the web application to have data pushed to it from the 
> > server (instead of polling, which was our short-term solution).  
> 
> Just curious.. why is this bad?
As long as you are not polling too often, it is not bad, but we were
trying to simulate real-time updates, which means _very_ frequent
polling, which would create an enormous load of the server (could even
be described as a DOS) in any environments with more than a few users.
> 
> 

> Perhaps the graphs become a flash/applet? But nothing else?
Based on feedback thus far, this is a definite possibility


> In a complicated production environment, you may not be allowed to 
> install this. Try asking out SOC guys to open up a wacky port for you.
> 
I was assuming this would be proxied in some way as with services like
xmpp in a webapp (like previous incarnations of rhx).

> Is the comet load on the server? I would assume not since there should 
> not be too many folks using the WUI.
> 
Yes, on the server, though the holding a connection open would also
affect the client.  And our assumption is that there could potentially
be _many_ people using the webapp, as mentioned above.

-j