Penetration Testing mailing list archives
Re: [PEN-TEST] Recourse Technologies -- info wanted
From: Peter Van Epp <vanepp () SFU CA>
Date: Tue, 3 Oct 2000 14:02:18 -0700
I've come to believe that this is more of a marketing tactic than an actual fact. I can believe that this would be true for an IDS with only a few signatures enabled, or one doing offline processing, but an IDS that is doing pattern matches on over 700 signatures in realtime, this is practically infeasible. Feel free to prove me wrong, but I've heard from
<snip> Or presumably help demonstrate you are likely right. On a PC class platform (i.e. without interleaved memory) a standard DIMM will do about 600 mbytes per second. NEC VC channel RAM gets that up to around a Gigabyte per second (i.e. 8 memory cycles available to process a full tilt boogie gig link). With VC channel ram that leaves you 6 or 7 memory cycles (depending on single cycle DMA from the card, and a cache system that doesn't need to copy the word across a kernel boundry in the OS which I find unlikely ...) to do your processing and run your OS in. If you have a high end Sun, SGI or other such (very uncheap!) machine you will get somewhat better performance (mostly due to interleaved RAM and faster busses) you will get a few more cycles but you will also get a likely inpractically large bill ... I expect even on a high end machine at inpractical cost you are looking at a couple of signatures max (i.e. impractically small in the real world, although we all know marketing people don't live there ...). Does anyone know what machine this performance is supposed to be achieved on (i.e. a several million dollar SGI origin2000 or a Cray)? Gig is going to be exciting (== expensive) to do IDS or firewalls on ... Peter Van Epp / Operations and Technical Support Simon Fraser University, Burnaby, B.C. Canada
Current thread:
- Re: [PEN-TEST] Recourse Technologies -- info wanted, (continued)
- Re: [PEN-TEST] Recourse Technologies -- info wanted Ben Rothke (Oct 03)
- Re: [PEN-TEST] Recourse Technologies -- info wanted Mark Teicher (Oct 03)
- Re: [PEN-TEST] Recourse Technologies -- info wanted Mark Teicher (Oct 03)
- Re: [PEN-TEST] Recourse Technologies -- info wanted Ryan Permeh (Oct 03)
- Re: [PEN-TEST] Recourse Technologies -- info wanted Ben Rothke (Oct 03)
- Re: [PEN-TEST] Recourse Technologies -- info wanted Erik Tayler (Oct 03)
- Re: [PEN-TEST] Recourse Technologies -- info wanted Ryan Permeh (Oct 03)
- Re: [PEN-TEST] Recourse Technologies -- info wanted subscribe (Oct 03)
- Re: [PEN-TEST] Recourse Technologies -- info wanted Mark Teicher (Oct 03)
- Re: [PEN-TEST] Recourse Technologies -- info wanted Paul Cardon (Oct 09)
- Re: [PEN-TEST] Recourse Technologies -- info wanted Oliver Friedrichs (Oct 03)
- Re: [PEN-TEST] Recourse Technologies -- info wanted Peter Van Epp (Oct 03)
- Re: [PEN-TEST] Recourse Technologies -- info wanted Mark Teicher (Oct 03)
- Re: [PEN-TEST] Recourse Technologies -- info wanted Andrew Lawton (Oct 04)
- Re: [PEN-TEST] Recourse Technologies -- info wanted Ben Rothke (Oct 03)