Well, monitoring products are rarely deemed "critical" systems within an organization. Those who do assign some level of criticality to their ESIM controls usually do so under fear of stiff regulatory compliance-imposed financial penalty. Security monitoring, as a service, is still attributed to a diminished level of importance within the grand sphere of operational information technology.
What about the loss of visibility into insider threat activities or the probing exercises of external attackers? Shouldn’t the mechanisms that provide this information be deemed critical? To the security-minded, yes. To the operationally minded, however, the loss of ESIM capabilities will almost always play second (or perhaps third or fourth) fiddle to the infrastructure providing availability for Internet and internetworked resources.
The bottom line is that the moving of packets and frames throughout an internetworked architecture will always take precedence over the reactive detection and reporting capabilities offered by an ESIM product. What most operationally minded people often fail to consider, however, is that if these products are not actively collecting, normalizing, correlating, and reporting on organizational data, the organization's ability to troubleshoot incidents is significantly diminished.
Hindsight is 20/20, and nowhere is this truer than in the world of ESIM monitoring. If your ESIM isn't operational, then you can’t collect the generated information, and you certainly can't view data related to an incident. Theologian Desiderius Erasmus of Rotterdam is often credited with the adage, "In the land of the blind, the one-eyed man is king," but how is such a philosophical statement relevant to the ESIM sector? Well, if the operations side of the organization doesn’t see your ESIM infrastructure as an important piece of the IT puzzle, then it is up to the security team to push for high-availability capable products. If the operations team can’t (won’t) elevate the criticality of your collection infrastructure, deploy redundant collectors (where possible) to ensure coverage.
If your log-generating devices are sending data following a single network path, ensure that they are sending to a secondary path in case the primary loses connectivity (sounds like a simple routing problem to me). If you can’t afford to dedicate physical hardware or vendor-backed appliances to do the job, perhaps you can explore the possibility of virtualizing some of the components for the sake of redundancy -- as most ESIM vendors are beginning to move toward software-only instances of their components. Ultimately, this will come down to a combined budget and technical feasibility issue.
However, planning for redundancy during the planning phase of an ESIM implementation is almost certainly better than months or years after the selected product is implemented.
Andrew Hay is senior analyst with The 451 Group's Enterprise Security Practice and is an author of three network security books. Follow him on Twitter: http://twitter.com/andrewsmhay.