Re: Changes in IDS Companies?

[Martin Roesch <roesch () sourcefire com>]

Network intrusion prevention systems are also relatively untested and 
still first generation.  The Hogwash wrapper for Snort (and the in-line 
mode being rolled into Snort) are both good technologies and intrusion 
prevention in general is a good idea, but the distance between "good 
idea" and a concept that's ready for larger market acceptance is a 
pretty wide gap.

One of the things that's been bothering me about the rush to build and 
deploy Network Intrusion Prevention Systems (NIPS) lately is the 
complete lack of discussion about the downsides of such technologies.  
My consternation falls into a couple categories that deal with the 
failure modes of NIPS and the political issues associated with 
deploying this kind of technology.

Most NIPS are built on the concepts pioneered by intrusion detection 
systems, protocol anomaly detection, signature-based analysis and 
traffic anomaly detection (port scans, etc).  Intrusion detection 
techniques are pretty well known for their applicability to specific 
problem areas, signature-based detection doesn't pick up attacks it 
doesn't know about, anomaly-based detection can't pick up signature 
based events (/cgi-bin/phf) very effectively.  The melding of these 
techniques is critical to providing good coverage from the perspective 
of a sensor designer, which is why Snort does signature and protocol 
anomaly detection (and several other tricks).  The problem is that *no* 
technology is capable of picking up every possible attack, a mix of 
technologies is often the best way to go to provide effective coverage 
of the security picture on a given network.

With this in mind, the basic question becomes "how do we know if our 
NIPS misses an attack?"  If the NIPS misses an attack,  we better have 
a pretty good NIDS/HIDS in place to let us know what happened.

How about failure modes of the technology itself?  It's been shown 
repeatedly in tests that NIDS technology can be notoriously unstable in 
a number of scenarios, what happens if that instability is translated 
to an in-line device?  We're either going to have a fail closed 
scenario (protected network is DoS'd)  or a fail open scenario in which 
the protected network becomes unprotected, possibly for a protracted 
period of time.  In the first scenario the failure mode will make 
itself apparent very rapidly, but in the second a NIDS/HIDS is going to 
be required to record and inform the security/admin staff about the 
problem as well as attacks during the lapse.

There's also the political battle of deploying another in-line 
technology in the network, etc. that will be fought anytime one of 
these is deployed, although I  think that fight will happen in the 
enterprise and not in the mid-tier market.

I'm an advocate of a layered solution.  Firewalls, NIDS/HIDS, 
authentication, crypto, etc. all continue to have their places on the 
network.  I think that host-based IPS will see quicker acceptance in 
the market than NIPS due to the lower "price of deployment/failure" 
associated with the host-based technologies, they're more like AV 
systems in their positioning as an end-host oriented security 
mechanism.  I think that there will definitely be convergence of the 
firewall and the NIDS, but I think it's early to declare these systems 
as the next generation, the political battle will have to be fought and 
the operational limitations of the technologies will have to be found 
before the final place of IPS in the network security "ecosystem" will 
be known.

     -Marty

--
Martin Roesch - Founder/CTO, Sourcefire Inc. - (410)290-1616
Sourcefire: Snort-based Enterprise Intrusion Detection Infrastructure
roesch () sourcefire com - http://www.sourcefire.com
Snort: Open Source Network IDS - http://www.snort.org

On Tuesday, October 15, 2002, at 04:45 AM, Avi Chesla wrote:

> I totally agree with you. Next generation IDS  ,also being called 
> Intrusion
> Prevention Systems or Perimeter Security devices are the next step in 
> the
> evolution of the Traditional Intrusion Detection Systems. Vendors such 
> as
> Intruvert, Tipping point ,  Vsecure Technologies , Lancope, Forescout ,
> TopLayer (Mitigator) etc, are example of some.
> All these vendors claim to have an Intrusion Prevention Systems which
> usually has some kinds of Adaptive capabilities, they do behavioral and
> protocol analysis and do not based on attack signature (most of them) 
> , they
> sit in-line (most of them), they mitigate attack without be depended in
> other products to do the blocking...
>
> Best Regards,
>
> Avi Chesla
> Director of Research
> Vsecure Technoliges, Inc.
> www.v-secure.com
>
> -----Original Message-----
> From: Samuel Cure [mailto:scure () netpierce net]
> Sent: Monday, October 14, 2002 10:54 PM
> To: focus-ids () securityfocus com
> Subject: Changes in IDS Companies?
>
>
> Just noticing some changes with some known IDS companies and wanted 
> some
> feedback from the community. Because Marcus Ranum left NFR earlier 
> this year
> and Ron Gula has left Enterasys Networks, I am questioning the future 
> of
> some early-on IDS companies. I mentioned some time ago that the IDS 
> market
> will eventually consolidate and it seems like things are moving in that
> direction.
>
>
> To further enforce my point, word on the street is TippingPoint is now
> seeking for someone to buy them out. Does anyone else have anything 
> that
> could help validate this or these types of trends in IDS companies?
>
>
>
> Thanks in advance!
>
> -------------------
> Samuel J. Cure
> Security Specialist
> NetPierce Security Services
> www.netpierce.net
> -------------------