[THIN] Re: ICA session bandwidth calculations

  • From: "Matt Kosht" <matt.kosht@xxxxxxxxx>
  • To: thin@xxxxxxxxxxxxx
  • Date: Thu, 8 Jun 2006 09:28:27 -0400

I am assuming the 27KB was bits not bytes, as 27 KBps = 216 Kbps...

I benchmarked a client/server app just recently.  Heads down data
entry in the application averaged only 26Kbps (no sound, 16 bit color,
128 bit encryption, seamless window) with v9.15.xxxx PNAgent.
Benchmarks of ESRI ArcGIS were closer to 33Kbps (probably because it
is so graphics intensive).

Printing can burst much higher as indicated but can be limited (using
the policy to control it in Citrix) without much notice by the user.
We limited GIS plotter users (who could plot 40-50MB drawings) to
30Kbps they didn't even notice.  Printers are so much slower than the
network it just sits in the spooler waiting to print vs. sitting in
the plotter's local memory waiting to print. Net impact to the user is
the same.

-Matt
On 6/2/06, Landin, Mark <Mark.Landin@xxxxxxxxxxxxxxxx> wrote:
I know the rule of thumb used to be 27KB or so for an ICA session. No
doubt things like high screen resolution, high color depth, and virtual
channels have changed that number somewhat.

Is there a number that is still used with some degree of confidence? If
not, can one try to calculate what a session would take up? If so, what
variables need to be considered?

For instance, I am thinking of deploying a GIS-like app, using 16-bit
color, 1280x1024 resolution, with no sound, from Windows 2003/MPS3 or
MPS 4, over an ICA 9 client. What could I expect the bandwidth
requirements for an ICA session running that app to be? Is there any way
to guestimate?
************************************************
For Archives, RSS, to Unsubscribe, Subscribe or
set Digest or Vacation mode use the below link:
//www.freelists.org/list/thin
************************************************

************************************************
For Archives, RSS, to Unsubscribe, Subscribe or set Digest or Vacation mode use the below link:
//www.freelists.org/list/thin
************************************************

Other related posts: