Welcome » IT Booklets » Information Security » Information Security Strategy » Architecture Considerations » Technology Design
A financial institution can significantly mitigate the risk of
security events by an appropriate technology design that provides
for effective network-level monitoring, limits an intruder's
ability to traverse the network, offers the minimum level of
services required for business needs, and is updated in a timely
manner to mitigate newly discovered vulnerabilities.
An effective means of accomplishing those goals is through the
use of security domains. A security domain is a part of the
system with its own policies and control mechanisms. Security
domains for a network are typically constructed from routing
controls and directories.
Domains constructed from routing controls may be bounded by
network perimeters with perimeter controls. The perimeters separate
what is not trusted from what may be trustworthy. The
perimeters serve as well-defined transition points between trust
areas where policy enforcement and monitoring takes place. An
example of such a domain is a demilitarized zone (DMZ), bounded by
a perimeter that controls access from outside and inside the
Domains constructed from directories may limit access to network
resources and applications based on role or function.
Directory-driven domains may allow access to different
network-driven domains. For example, a network management
domain may use the same cabling and network interface cards as
other domains, allow access to all computing devices in all
domains, but limit the allowed access based on the user's role or
The selection of where to put which control is a function of the
risk assessment. Institutions generally should establish
defenses that address the network and application layers at
external connections, whether from the Internet or service
providers. Internally, perimeters can be established at higher-risk
security domains, such as wire transfer, and to segregate at a
network level those areas of the institution that work with
customer information from other areas. Internal perimeters
also may be used to create security domains based on geography or
other logical or physical separations.
Hosts may also include security perimeters. Those perimeters are
enforced through authorizations for users and programs. The
authorizations can be a part of applications, the file system, and
the operating system.