IDS mailing list archives
Re: Asymmetric traffic/topology
From: Adam Powers <apowers () lancope com>
Date: Fri, 09 Nov 2007 17:47:47 -0500
If when you say "Behavior-based System" you're referencing any of the NetFlow-based products then you're not quite right. Behavior-based NetFlow products overcome the asymmetry problem by reassembling flows from multiple routers along disparate data paths. If we have a situation such as: Client -> R1 -> R2 -> R3 -> Server Client <- R1 <- R4 <- R3 <- Server ..in this case the path is R1 to R2 to R3 in one direction and R3 to R4 to R1 on the return path. Classic asymmetric L3 routing. NetFlow gets around this problem because R1, R2, R3, and R4 are all sending their NetFlow exports to a single collector where the unidirectional flows are merged and processed into a "bi-flow" that would look more like... Client <-> R1 <-> R2,R4 <-> R3 <-> Server Analysis can then be performed on the bidirectional flow without fear of asymmetry issues. All that's required is that you enable NetFlow on the correct devices (the more the merrier IMO). On 11/8/07 6:06 PM, "Jeremy Bennett" <jeremy () deities org> wrote:
First there are three types of asymmetry in a network that can cause problems for some times of IPS devices. 1. Connection-level asymmetry: This is the case where a given TCP connection (up and down stream) is on a single network path but a separate, identical connection may follow a different path. This is very common and can cause problems for behavioral systems. 2. Flow-level asymmetry: This is the case where the upstream and downstream flows in a TCP connection may follow different paths. This can cause problems for behavioral systems and stateful packet- inspection. 3. Packet-level asymmetry: This is the case packets within a flow may be following different routes in a network. This can cause problems for any IPS except for the most basic packet-filter. Now in my experience, #1 is very common in medium to large enterprises that have built for scalability and redundancy. #2 is common in load-balanced server farms. #3 is not extremely common but does appear in some instances of a hot-hot redundancy deployment. -J On Nov 7, 2007, at 4:42 PM, snort user wrote:Greetings. I am sure that most of you know about the asymmetric traffic/topology problem in relevance to IDS/IPS systems. ( By Asymmetric traffic/topology, I mean the case where client to server packets traverse a different path in your network compared to server to client packets. Hence the IDS/IPS see only one side of the conversation) I am trying to find out how wide this problem really is? Is it commonly seen in large / enterprise networks ? Any input is welcome. Thanks ---------------------------------------------------------------------- -- Test Your IDS Is your IDS deployed correctly? Find out quickly and easily by testing it with real-world attacks from CORE IMPACT. Go to http://www.coresecurity.com/index.php5? module=Form&action=impact&campaign=intro_sfw to learn more. ---------------------------------------------------------------------- -------------------------------------------------------------------------- Test Your IDS Is your IDS deployed correctly? Find out quickly and easily by testing it with real-world attacks from CORE IMPACT. Go to http://www.coresecurity.com/index.php5?module=Form&action=impact&campaign=intr o_sfw to learn more. ------------------------------------------------------------------------
-- Adam Powers Chief Technology Officer Lancope, Inc. c. 678.725.1028 f. 678.302.8744 e. adam () lancope com ------------------------------------------------------------------------ Test Your IDS Is your IDS deployed correctly? Find out quickly and easily by testing it with real-world attacks from CORE IMPACT. Go to http://www.coresecurity.com/index.php5?module=Form&action=impact&campaign=intro_sfw to learn more. ------------------------------------------------------------------------
Current thread:
- Asymmetric traffic/topology snort user (Nov 08)
- RE: Asymmetric traffic/topology Bergen, Matt (Nov 09)
- RE: Asymmetric traffic/topology Srinivasa Addepalli (Nov 09)
- Re: Asymmetric traffic/topology Jeremy Bennett (Nov 09)
- Re: Asymmetric traffic/topology snort user (Nov 09)
- Re: Asymmetric traffic/topology Jeremy Bennett (Nov 09)
- Re: Asymmetric traffic/topology Ravi Chunduru (Nov 09)
- Re: Asymmetric traffic/topology Adam Powers (Nov 13)
- Re: Asymmetric traffic/topology Jeremy Bennett (Nov 13)
- Re: Asymmetric traffic/topology Roland Dobbins (Nov 14)
- RE: Asymmetric traffic/topology Nelson Brito (Nov 27)
- Re: Asymmetric traffic/topology snort user (Nov 09)