Re: Interpretation of measurements

["Robert MacDonald" <rmacdonald () gfs com>]

Tamas,

Our system is handling hundreds of connections, so I'm
not a good source of true usability. I find this is one of
those hard to answer questions, because of all the
variables involved. A couple come to mind.

Connections per second that a vendor tell in the marketing
material can be (is) misleading. There has been some
talk about how the tests are run. Some vendors are only
telling the number of connects created and torn down,
without sending data across the connect. This gives a
false value for what the system can handle - which is not
real world IMO.

Another big variable is encryption. If your encrypting, then
I highly doubt you'll get that many connects, unless you
have a _very_ powerful system that can handle the CPU
intensive crunching needed. This is one of the biggest
reason the online stores et'al don't do encryption
until you go to buy.

There are others who do have experience in this area,
and hopefully they will post their thoughts on this.

Robert

- -
Robert P. MacDonald, Network Engineer
e-Business Infrastructure
G o r d o n   F o o d    S e r v i c e
Voice: +1.616.261.7987 email: rmacdonald () gfs com

> > > Forjan Tamas <tamas () 2fkft com> 8/15/00 7:35:51 AM >>>
> Dear Wizards,
>
> many firewall vendors list some metrics about their products that they
> feel will give an idea of the performance of the products. Such a
> measurement is the connections/sec value.
>
> I don't have too much experience with high-traffic sites. Those of you who
> have, how would you interpret "1000/2000 connections/sec" into real-life
> usability (e.g. number of users)?
>
> I am also interested in hearing what connections/sec numbers do your
> firewalls reach at how many users.
>
> Thanks a lot,
>
> -- 
> Tamas FORJAN
> technical support
> http://www.2f.hu/ 



_______________________________________________
Firewall-wizards mailing list
Firewall-wizards () nfr net
http://www.nfr.net/mailman/listinfo/firewall-wizards