Snort mailing list archives
Re: Snort performance concerns
From: sekure <sekure () gmail com>
Date: Fri, 30 Sep 2005 10:40:57 -0400
First place i'd look is disabling rules you don't need. Not sure if you've done that already, but many people are running with much larger rulesets than what they need. I was seeing the same thing you were, for example, certain sensors are handling near 100Mbps with minimal drops, while others are dropping packets at > 3Mbps. After disabling some unnecessary rulesets the CPU utilization went way down and i haven't seen any issues since. I've seen the highest performance hit with the web-client rules for some reason. So disabling try that first. On 9/30/05, Larry Wichman <larrywichman () yahoo com> wrote:
I enabled Performance Monitor on my sensors and I have some concerns after looking at some of the performance stats. First, I have three sensors, two of which average 96mb/sec of traffic and the dropped packets percentage average is about 10% (proc and memory utilization are high, as expected). I have a third sensor that sees an average of about 5mb/sec and has the same amount of dropped packets, memory and proc utilization are minimal. I have implemented all the suggested optimizations (I think), patched Libpcap, etc….I can understand that there would be some dropped packets when the traffic is at a high, continuous load, but the third sensor with the same amount of dropped packets with only a fraction of the traffic concerns me. I am thinking about upgrading the hardware (faster proc, bus speeds, etc…), but I might be wasting money if the stats are the same. Does anyone have any input as to what is causing the dropped packets? Also, my boss told me to start evaluating commercial products. My first choice would be Sourcfire, I really do like working with Snort, but I need whatever product I choose to be able to handle the amount of traffic that we have. I would greatly appreciate any input on this. Cheers. Larry
Current thread:
- Snort performance concerns Larry Wichman (Sep 30)
- Re: Snort performance concerns Joel Esler (Sep 30)
- Re: Snort performance concerns Larry Wichman (Sep 30)
- RE: Snort performance concerns Jeff Dell (Sep 30)
- Re: Snort performance concerns Joel Esler (Sep 30)
- Re: Snort performance concerns Larry Wichman (Sep 30)
- Re: Snort performance concerns sekure (Sep 30)
- RE: Snort performance concerns Jeff Dell (Sep 30)
- <Possible follow-ups>
- RE: Snort performance concerns Joshua Berry (Sep 30)
- RE: Snort performance concerns Richard Bejtlich (Sep 30)
- Re: Snort performance concerns Joel Esler (Sep 30)