What I would suggest is opening the process viewer and seeing what process 
allocates the memory. WebORB for .NET memory footprint is 15MB. However, if the 
code you host in there allocates a byte array for 5.6GB, that could explain the 
problem..

Mark

--- In [email protected], "valdhor" <valdhorli...@...> wrote:
>
> I do not believe the Flex SWF's are taking up the memory (There is no way for 
> them to do so).
> 
> It is more likely the language you are using with Flash remoting and WebORB. 
> Is it Java? I can see an improperly written Java class reserving large 
> amounts of mamory. What is the JVM settings?
> 
> --- In [email protected], "kevin_ketterer" <kevinkett@> wrote:
> >
> > Sorry, the Flash SWF did NOT reserve any memory... so only the two Flex 
> > SWFs did.
> > 
> > --- In [email protected], "kevin_ketterer" <kevinkett@> wrote:
> > >
> > > We have a Flex-based shopping cart applicatiion, using Cairngorm, WebORB, 
> > > and ASP.NET. It's running on IIS 7/Windows 2008 server. The release 
> > > version of the SWF is about 1.4MB in size. This is the only application 
> > > hosted on this server.
> > > 
> > > Our server team tells us that when a client launches the application, it 
> > > reserves about 5.6GB of virtual memory, including 3GB of RAM. It only 
> > > used the memory it neede, however. Since this is a new installation for 
> > > the server team, we ran some comparison tests to see what "normal" is. 
> > > Each of the following apps were put in their own application pool:
> > > 
> > > * An application with a plain ASP.NET (HTML) page did not reserve the 
> > > memory.
> > > 
> > > * A small (~150k) Flex SWF in an HTML page reserved 5.6GB, but again used 
> > > much less.
> > > 
> > > * A 4k Flash SWF in an HTML page reserved 3GB.
> > > 
> > > Although this behavior is not causing an immediate problem, we are 
> > > concerned that this could become a bigger issue down the road. We would 
> > > also like to understand how a SWF that runs on the client might affect 
> > > server virtual memory.
> > > 
> > > I'll appreciate any thoughts you might have on this issue.
> > > 
> > > Thank you,
> > > Kevin
> > >
> >
>


Reply via email to