Untitled Document
There is no shortage of available information on security products. Fortunately,
the forces of natural selection effectively rid the market of inferior and uneconomical
offerings. Information security professionals are able to select from a pool
of best-of-breed products with which they can feel reasonably confident and
secure. But there is often an ironic consequence to this: the more fit a security
product is perceived to be, the more likely it is to recklessly embolden those
whom it is employed to protect, having the effect of instilling a deleterious
illusion of security.
Still, this a small problem compared with the fact that attacks evolve according
to the same forces driving the advancement of defensive countermeasures. The
fittest attacks adapt to their adversaries, becoming increasingly stealthy and
subtle, both in their delivery and the perceptibility of their payload.
Detection at the time of occurrence relies on intrusion detection/prevention
systems that have ever-reduced visibility into increasingly covert attacks.
Security Event/Information Management (SIEM) platforms can only report on and
respond to the specific events for which they have been configured. Log aggregation,
analysis, and correlation tools can only act on the specific meta-information
about the set of events that they have been programmed to recognize. Application
layer gateways, proxies, and their derivatives can only operate on the well-known
protocols, procedures, and methods they were written to handle. Deterministic
(pattern- or signature-based) methods of detection have mounting difficulty
dealing not only with intentional obfuscations, but also with the inevitable
window of exposure that exists between the introduction of an attack and the
development and deployment of antidotal signatures.
In response, some defensive systems are moving toward a cocktail of deterministic
and heuristic (behavioral/anomaly based) methods of detection. Unfortunately,
the latter, because of its deficient certitude relative to the former, often
cannot be employed with sufficient aggressiveness to achieve comparably material
effectiveness, lest it introduce insufferable false-positives. But these technologies
will mature. And naturally, once these hybrid systems become sufficiently pervasive,
the survival of the attacks will come to depend on their fitness at simultaneously
impersonating "normal" behavior while minimizing detectability. The
arms race will continue.
-1-