Introduction
Network penetration testing—using tools and processes to scan the network environment for vulnerabilities—helps refine an enterprise’s security policy,identify vulnerabilities, and ensure that the security implementation actually provides the protection that the enterprise requires and expects.Regularly performing penetration tests helps enterprises uncover network security
weaknesses that can lead to data or equipment being
compromised or destroyed by exploits (attacks on a network,
usually by "exploiting" avulnerability of the system),
Trojans (viruses), denial of service attacks, and other intrusions.
Testing also exposes vulnerabilities that may be introduced
by patches and updates or by misconfigurations on servers, routers,
and firewalls.Penetration Testing Overview
The overall objective of penetration testing is to discover areas of the
enterprise network where an intruder can exploit security vulnerabilities.
Different types of penetration testing are necessary for different types of
network devices. For example, a penetration test of a firewall is different
from a penetration test of a typical user’s machine. Even a penetration test
of devices in the DMZ (demilitarized zone) is different from performing a
scan to see if network penetration is possible.The type of penetration
test should be weighed against the value of the data on the machine being
tested and the need for connectivity to a given service.
The penetration testing process has three primary components:
• Defining the scope
• Performing the penetration test
• Reporting and delivering results
Step 1: Defining the Scope
Before a penetration test can be launched, the enterprise must define the
scope of the testing.This step includes determining the extent of testing,
what will be tested, from where it will be tested, and by whom.
Full-Scale vs.Targeted Testing
An enterprise must decide whether to conduct a full-scale test of the
entire network or to target specific devices, such as the firewall. It is usually
best to do both in order to determine the level of exposure to the public
infrastructure, as well as the security of individual targets. For example,
firewall policies are often written to allow certain services to pass through
them.The security for those services is placed on the device performing
those services and not at the firewall.Therefore, it is necessary to test the
security of those devices as well as the firewall. Some of the specific targets
that should be considered for penetration testing are firewalls, routers,Web
servers, mail servers, FTP servers, and DNS servers.
Devices, Systems, and Passwords
In defining the scope of the project, the enterprise must also decide on the
range of testing. For example, is it looking only for vulnerabilities that could
lead to a compromise of a device, or is it also looking for susceptibility to
denial of service attacks? In addition, the enterprise must decide whether
it will allow its password file to be hacked by the security team to test
its users’ choice of passwords, and whether it will subject its devices to
password grinding across the network.
Remote vs. Local Testing
Next, the enterprise must decide whether the testing will be performed
from a remote location across the Internet or onsite via the local network.
This decision is dictated to a large degree by the targets that are selected for
testing and by the current security implementations. For example, a remote
test of a machine behind a firewall that hides network address translation for
Internet access will fail if the firewall appropriately prevents access to the
machine. However, testing the same firewall to see if it will protect users’
computers from a remote scan will be successful.
In-House vs. Outsourced Testing
After the scope of the testing has been determined, the IT team must
decide whether to use in-house resources to perform the testing or to hire
outside consultants. In-house testing should be chosen only if an enterprise
lacks the funds to hire outside consultants, or if the data is so sensitive
that no one outside the company should view it. In all other cases, hiring
outside consultants is recommended. Outside security consultants are highly
trained and have worked with hundreds of different networks, bringing
specific expertise and broad experience to the testing process. In addition,
they help ensure an unbiased and complete testing procedure. Security
consultants continuously research new vulnerabilities, invest in and
understand the latest security testing hardware and software, recommend
solutions for resolving problems, and provide additional personnel for the
testing process. Enterprises can leverage the experience and resources of
outside security consultants to help ensure thorough, properly executed
penetration tests.
Step 2: Performing the Penetration Test
Proper methodology is essential to the success of the penetration test. It
involves gathering information and then testing the target environment.
The testing process begins with gathering as much information as possible
about the network architecture, topology, hardware, and software in order
to find all security vulnerabilities. Researching public information such as
Whois records, SEC filings, business news articles, patents, and trademarks
not only provides security engineers with background information, but also
gives insight into what information hackers can use to find vulnerabilities.
Tools such as ping, traceroute, and nslookup can be used to retrieve
information from the target environment and help determine network
topology, Internet provider, and architecture.Tools such as port scanners,
NMAP, SNMPC, and NAT help determine hardware, operating systems,
patch levels, and services running on each target device.Three Level of Testing Services:
Common Vulnerability Assessment (CVA)
The CVA is a remote security assessment that focuses on the services
that are most commonly misconfigured by personnel and are most
commonly exploited by intruders. It also focuses on the most probable
means of unauthorized access. A professional security engineer not only
interprets the scanner output but also creates an executive summary and
recommendations report.
Secure Device Assessment (SDA)
The SDA is an on-location device configuration assessment that includes
architectural review of device deployment, operating system configuration,
and device and policy configuration.This assessment is similar to an audit,
except that it includes scanning services, when necessary.
Secure Exploit Assessment (SEA)
This penetration study encompasses all aspects of the CVA and also includes
the following features: additional vulnerability research, DNS auditing, full
enumeration including NetBios and Windows NT- and Unix-specific
issues, penetration attempts with multi-stage attacks, and custom attack
methodologies. Additional options include "brute force" password cracking
and grinding, blind scanning (attacker perspective),"war dialing," and testing
for denial of service attacks and social engineering (manipulating users to
obtain confidential information such as passwords).
