Re: Firewalling performance

1998-04-06 Thread James Youngman
""Eric" == "Eric L Green [EMAIL PROTECTED]" [EMAIL PROTECTED] writes: [Linux may not cope with the load of a T3] According to tytso, you can drive two serial lines at 115Kbps simultaneously on a 386 under Linux. That's at least one interrupt per sixteen bytes, usually more. With networks

Re: Firewalling performance

1998-04-06 Thread Chris Evans
On 6 Apr 1998, James Youngman wrote: ""Eric" == "Eric L Green [EMAIL PROTECTED]" [EMAIL PROTECTED] writes: [Linux may not cope with the load of a T3] [re: firewalling loads] As a real datapoint, our Linux box doesn't bad an eyelid with stacks of firewall rules under full ethernet load.

Re: Firewalling performance

1998-04-04 Thread Jason Costomiris
On Fri, Apr 03, 1998 at 03:18:05PM -0600, Eric L. Green [EMAIL PROTECTED] wrote: : On Wed, 1 Apr 1998, Fred Leeflang wrote: : In the company I work for, we're considering setting up a Linux : firewall. I do have some experience with it, know how to create firewall : rules and such, but

Re: Firewalling performance

1998-04-03 Thread Eric L. Green [EMAIL PROTECTED]
On Wed, 1 Apr 1998, Fred Leeflang wrote: In the company I work for, we're considering setting up a Linux firewall. I do have some experience with it, know how to create firewall rules and such, but I've never been in the opportunity to see how well Linux holds up as a firewall under high

Re: Firewalling performance

1998-04-03 Thread William T Wilson
On Fri, 3 Apr 1998, Eric L. Green [EMAIL PROTECTED] wrote: Personally I would not use Linux to do NAT and packet filtering. Most Linux does have a big whopping advantage in two cases: 1) You want to route between ethernet segments, 2) You want to run a proxy on the same host, and 3) You have

Re: Firewalling performance

1998-04-03 Thread Chris Frost
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 On Fri, 3 Apr 1998, William T Wilson wrote: a Pentium-200 inside. Based on the output of concurrently running benchmarks (the rc5 cracker) compared to an unloaded P166, the routing is only using something like 15% of the CPU on the P150. Also,

Re: Firewalling performance

1998-04-03 Thread William T Wilson
On Fri, 3 Apr 1998, Chris Frost wrote: Also, the 2.1.x development kernels have *significantly* faster networking, so 2.2.x will be much faster than 2.0.x (obviously I guess ;-) I did not mention that I am using the 2.1 kernel to do this. But how much faster is the networking going to get?

Firewalling performance

1998-04-01 Thread Fred Leeflang
Hi, In the company I work for, we're considering setting up a Linux firewall. I do have some experience with it, know how to create firewall rules and such, but I've never been in the opportunity to see how well Linux holds up as a firewall under high loads. The system we're thinking about