Network Access Control
The challenge facing modern enterprise network administrators is how to
handle the different types of users and their devices who are entering
the network remotely as well as within the enterprise. How is
an enterprise to manage the identities of users, devices, locations,
applications, and at the same time perform host integrity
checks? How do enterprises create a single point of policy
enforcement and eliminate the existing security policy silos?
The answer lies in using layered identity based enterprise
security.
First Layer - Identity Registration
The first layer is to do identity checks on workers who will be
enterring the enterprise electronically. Remembering the adage that 80%
of attacks come from within the enterprise, it's important to screen
out people with a criminal past. As well, you should be doing
the
same for your janitors (review the paper "
Why
your use of id and password is likely a joke").
Second Layer - User Training
Are you currently doing any ongoing training with your users about
email, trojans, phishing and pharming attacks? If not, why
not?
One of the easiest ways to break into an enterprise is to use
email with a MS Office attachment containing malware. When
the
user clicks on it, the malware is released behind the firewall. At the
minimum, you should be doing some kind of regular training to remind
workers to be very vigilent on what they click on in an email.
Third Layer - Device Security
The third layer is the actual device or appliance the user is coming in
from. A modern enterprise uses network hardware/software
appliances that intercept all enterprise network requests. The
appliance then determines the integrity of the host by scanning the
user's PC, laptop, wireless or handheld device to see if the device
meets acceptable security configurations.Devices that don't
meet the configuration are immediately transferred into a quarantine
area where remediation is offered to the device.
Fourth Layer - User's Identity Initial Authentication
The next security layer is to take the user's identity and
determine what network access is applicable using the users' access
rights. This layer of security interacts with the enterprise
identity authentication, authorization and audit security
engines to determine the authentication required. Normally, the
hardware device and/or the enterprise single sign on system challenge
the user for the
required authentication. The authentication response is then normally
checked against the enterprise LDAP (Lightweight Directory Access
Protocol). This often involves multi-factor
authentication using user id's, passwords, digital certificates,
security tokens, smart cards and biometrics.
Fifth Layer - Quick Provisioning and De-provisioning
When
a user
no longer requires access to your enterprise, an application, building,
room,
network, etc., how long does it take until they are de-provisioned? Many enterprises have very
weak to poor
provisioning and de-provisioning processes.
In today’s age, this puts the enterprise at
greater risk,
since a user
who is gone may still have access to the enterprise.
Put in place the infrastructure to quickly
add, adjust or remove someone from having physical or electronic access
to your
enterprise
Sixth Layer - Stronger Authentication
As
the
enterprise risk rises for networks, applications and information
access, so too
must the layers of authentication strength.
The financial system, payroll and payables are all higher
risk. So too are
users who hold super-user
privileges like senior network administrators.
For
all of
the medium and higher risk applications, your enterprise should be
using a
graded series of stronger authentication.
For instance, low to medium risk might be addressed by the
user
providing their id, password and a digital certificate. Medium
risk
should be addressed by the user providing things like a secureID token
along
with their id and a password. Medium
to
high risk should be addressed by the user providing something like a
smart
card, a secure id token, a biometric and a second unique password.
Seventh Layer - Re-Imaging Network Operating Systems
When your enterprise gets hit by a successful rootkit attack in the
future, what are you going to do? Currently, Microsoft
recommends
re-imaging all infected computers on the network. This could
effectively slow down the enterprise and stop most IT department
activities as personnel scramble to deal with the rootkit.
Avoid
this by having a plan in place that will quickly, at a low cost,
re-image the network and stop the infection from spreading or worse,
re-appearing.
Eighth Layer - Transaction Authentication
You must assume that all the previous layers have been breached.
How do you protect your enterprise crown jewels.
You need
to deploy transaction authentication for your high risk information,
applications and networks.
In
transaction authentication, software watches the following:
- IP address
being used by the user
- Geo-location
of the user
- Time of
day the event is occurring
- Historical
user pattern
- Computer
hardware the user is using
If
any of
these criteria are different than expected, even with a successful
authentication, the transaction authentication software will start
alarm bells
ringing in the enterprise. This may result in:
- The user
being asked all sorts of personal questions to verify it is really them
- Security
or business managers being paged in real time
- The event,
process or transaction refused
Ninth Layer
The final
layer of defense is the ability to quickly go back in minutes, hours,
days,
weeks or months to find out every network, application, information
resource,
type of device used, building and room the identity used at a specific
point in
time. All too
often, the audit data is
very hard to interpret from an application audit trail and worse, hard
to
integrate with other audit data from other applications.
Get
a team
organized on trying to provide an end to end user audit that is quickly
available, easy to access and use.
This
will help you find out where problems and breaches occurred and then
prepare
remedial action.
CONCLUSION
Controlling network access is therefore heavily dependant upon the
underlying identity infrastructure since this information is used to
authenticate and authorize the user from enabling them to access
network, systems, applications and information. A layered network
access control security strategy is required in order to adequately
protect
the enterprise based on its own internal risk assessment.
Network
Access
Control Security Strategy 2006 - A white paper
Authentication - Enterprise Security
Password
Authentication
Single
Sign On Authentication Access
Control Authentication Authentication-Enterprise
Security Authentication
Strength Authentication
Transaction
Authentication
Management User
Authentication Authentication
Federation Biometric
Authentication PKI
Authentication Token
Authentication Wireless
Authentication Document
Authentication
Authentication - Outsourcing