Hi Jozsef Kadlecsik,
I never tested a high performance network but my tests with NPtcp (NetPipe)
and tcpblast give me the same result as in production.
You must have more high speed machines on both site of the packet filter
because the filter performance are higher than create or receive the traffic.
I use 4 machines on both sites to bring the packetfilter to his upper
limit.
Eddy
you wrote:
> Hello,
>
> I would like to ask your opinion on what kind of PC hardware (CPU,
> motherboard, RAM, FastEthernet card) would be suitable for
> the following task:
>
> - - stateful packet filtering on Linux
> - - 155Mbit/s bandwidth
>
> I know the full duplex FastEthernet is not equivalent to STM1, but
> unfortunately the Gigabit Ethernet or ATM are not possible alternatives
> on the firewall machine. Therefore the good FastEthernet cards are quite
> important.
>
> Also, what is the best way to test the throughput of a hardware
> doing stateful packet firewall in a test environment? How could one
> generate "real" TCP/UDP traffic? Is there a better way than to setup
> thousands of echo services "inside" and connect them from "outside",
> for example? How can one measure the bottlenecks on the firewall, i.e
> the limit of what hardware component (slow CPU, slow bus, not enough RAM,
> slow/scrappy Ethernet card, etc) is reached at the current traffic?
> Is there any ready to use software available for the testing?
>
> Thank you any help,
> Jozsef Kadlecsik
--
-----------------------------------------------------------------------
Edwin Cremer Monday, February 28, 2000 2714 Std. ohne Absturz
[EMAIL PROTECTED] PGP-Key-ID: 3AF75CD9
PGP-Key-fingerprint: 22 D7 06 43 3E 04 3F FA D4 7A 67 EA 6B 70 60 25
-----------------------------------------------------------------------
It's deja vu all over again.
-
[To unsubscribe, send mail to [EMAIL PROTECTED] with
"unsubscribe firewalls" in the body of the message.]
