Cyber Criminals Gaining
Support from Overseas Operations By Gordon Smith (July 2009)
Download
PDF of article at http://www.canaudit.com/Perspectives/Volume10-Issue1.pdf
I am very concerned about
recent cyber incidents. There has definitely been an uptick
in incidents,
with new incidents reported daily. Even more interesting, the
increase is in organized attacks and scams rather than single
hacker-type attacks. These broad-based attacks have been around
for years, increasing with the invention of bot armies (PCs
that are compromised, backdoored, and available to attack sites
in concert). These attacks have reached a fever pitch as new
scams have been developed and executed against poorly controlled
American businesses. The most detailed report I have seen regarding
one of these newer attacks appeared in the Washington Post
on July 2, 2009 (see http://voices.washingtonpost.com/securityfix/2009/07/an_odyssey_of_fraud_part_ii.html).
The article details how
cyber criminals engineered a scam to create wire transfers
from a Bullitt County, Kentucky bank
account. The cyber thieves used spyware to capture the accounts
and passwords. They then changed the email address used by
the bank to confirm wire transfers. After this was complete,
they proceeded to steal $415,000. The interesting part of this
case is that they “hired” unsuspecting people,
using online employment sites, to receive the fraudulent transfer
and forward it to overseas banks. The Washington Post article
does an excellent job of describing the mechanism for this
fraud. It is, to say the least, quite a bold scheme. My concern
as an auditor is that this fraud indicates the need for additional
client-side controls, which are currently missing from many
applications. In addition, we need new controls such as automated
scrutiny of online transactions.
This scheme was able to
work because the client machine was trojaned. Popular opinion
is
that the client must absorb the
risk. In this case, I believe that the bank, First Federal
Savings Bank, could be impacted by the bad publicity related
to the incident (see http://www.fox41.com/Global/story.asp?s=10627534).
The reputational risk is such that we now need to reach out
to our clients to ensure that their machines and
the applications on their machines are safe to connect to our
network.
I would like to dig a
little deeper. Let us look to other areas where new fraud puts
organizations at risk. My
first
thought is the growing number of people who work from home.
They may work from home full-time, one or two days a week,
or sporadically. They may use corporate or personal machines.
If they are using corporate machines, one would expect the
machines are protected with anti-malware software to prevent
against viruses, trojans and other malicious software. I
am also very concerned about staff using their home computers
to login to their work email. I know that many people do
not
have anti-virus/malware software on their machines. Others
have installed the software but stopped paying for the updates
necessary to keep it current. As a result, their machines
are at risk. Each time they visit a web site or open an email
they
may inadvertently load a keystroke logger onto their machines.
When they log into the VPN or webmail application, their
login credentials may be compromised.
If staff members are logging
into a VPN, remember that they may have access to the internal
network. If their home machine
is backdoored, trojaned, or has an unauthorized keystroke
logger installed, the cyber criminals may be able to login
to the
VPN and have access to the internal network and the devices,
files and databases within it. Once they have the credentials
to login to the VPN, criminals can log themselves in to capture
data or take control of corporate machines for use as part
of a broad-based attack on our national information infrastructure
(covered later in this article).
Some of our clients require
two-factor authentication. Those using tokens and other external
devices, such as smart cards,
to login should still be protected if the account and password
is compromised. Those using digital certificates may not
be so lucky. Once the cyber thief has access to a machine,
they
may be able to escalate their rights in order to steal the
digital certificate or cookies
required to complete a successful login. In previous articles
I have described how to escalate rights and easily take control
of databases (see http://www.canaudit.com/articles.html).
Extend these concepts to home PCs that have been compromised
and we
have an Information Technology Perfect Storm. Your security
is defeated and your data and databases lie exposed to the
hackers.
Home workers are not the
only concern. There are also contractors who may login remotely
through the VPN. This could
be from
their office in Canton, Ohio or Bangalore, India. If their
machines are compromised and credentials stolen, then the cyber
criminals will have access to program and data files the contractors
have. Again, they may be able to use this access to escalate
their rights up to and including the domain level. They could
also compromise the Storage Area Network (SAN) to gain complete
access to all of your organization’s critical information.
Some of the SANs have known default passwords which we have
used on several client audits. Yes, they work. The warning
here is that if someone is able to get into the internal network
using compromised credentials, LogMeIn, or a similar product,
they can easily launch a SAN attack.
Now that the risk is established,
lets look to controls. The first and most important is to ensure
all machines that connect
to your organization’s network, both internally and externally,
use anti-virus and anti-spam software. The software should
be updated as soon as updates are made available. It is no
longer acceptable to wait several days or weeks to implement
vendor updates to the malware signatures. This may be difficult
for your clients. If they do not have the software, you can
suggest they get it. However, there is not a lot of leverage
here as customers generally do not want to buy additional software
and vendors do not want to give the impression that their web
sites may not be secure.
For customers that have
high-value or high-volume accounts, you may want to provide
the software
and installation assistance
to ensure that the anti-malware is in place and properly installed.
This is only a slight modification of the old giveaway programs
used by some banks: “Open a new account and get a new
toaster!” Now it is: “Do business with us online
and we will provide security software to protect your transactions.”
Contractors
should also be required to ensure that their machines have
up-to-date anti-malware software and current malware signatures.
To enforce this, you should execute your right-to-audit clause
on some of your vendors each year. I suggest taking some
of
the largest, plus one or two of the smallest, and subjecting
them to an onsite audit. Whether malware comes in through a
contractor providing 300 engineers for your project or through
a sole proprietor contractor, the damage to your firm can be
the same. Another good control is to subject the contractor
to a security test prior to awarding them a contract. This
could be in several forms such as a self-completed checklist
or an onsite audit. While this may seem far-out by today’s
standards, in five years it will be the norm as malware incidents
drive audit risk and future audits.
Do not forget to audit
your offshore contractors and outsourcers as well. Every one
or
two years, the IT auditors should visit
the foreign contractors to ensure that all contract terms,
corporate policies and security requirements are met or exceeded.
I am very concerned that authentication credentials may be
shared by overseas contractors, particularly larger firms that “follow
the sun”. These contractors use their various offices
around the world to support your IT requirements. As one office
closes, your support may transfer to another office. The login
credentials may move from one office to the next, creating
a serious security issue.
It is also important to
monitor the patterns of those who connect to your network.
If Gordon Smith
normally does two
transactions a week, and today he has performed seven transactions,
it is possible that the extra activity is fraudulent. It would
be prudent to call Mr. Smith to confirm the transactions. If
any of Mr. Smith’s personal data changes, particularly
email address or telephone number, it could be a signal that
his credentials have been compromised. The email and phone
numbers are used to contact a client when suspicious activity
occurs. Watch for changes to the contact information. These
changes should be confirmed by email to the old and new email
addresses and possibly by phone to the old phone number as
well. My stockbroker puts a hold on my online transactions
after changes have been made to my personal information. I
have to perform transactions over the phone until the confirmation
window is concluded. As soon as a change is made to my information,
they send out an email and a letter. The confirmation window
is the mail time of the letter plus two days to ensure that
I have read it. Although inconvenient, this is a good control.
Application
controls must also be implemented. Daily and weekly transaction
limits can be useful in preventing blatant procurement
or EFT frauds. Variances in shipping addresses can be useful
to detect diversion of goods frauds. Mailed and emailed confirmations
can serve as detective controls provided changes to personal
or billing information is properly confirmed as previously
mentioned.
The last suggested control
I will discuss relates to a security breach involving personal
health information (PHI). Medical records were recently compromised
by what was termed a “virus”. The “virus” mentioned
in the article is called Coreflood (see http://www.healthcareitnews.com/news/virus-blamed-ehr-breach-canada).
Several times a day, machines connected to your network should
be scanned for malware. Ensure that your security folks
test for all known spyware and malware. In some cases, they
will balk at this suggestion. I consider this the same as car
seat belts when they first came out. Many people did not wear
them, as they did not see the benefit. Even I balked at wearing
one until the police started to enforce it. Now we know that
seat belts really do save lives. Let me tell you, actively
searching for spy and malware will lead to early detection
of a potential security breach and mitigate the damage that
can be done.
The second item I want
to cover in this article is the ability for nations, individuals,
groups and conglomerates
to paralyze
a website. Earlier this month, the alleged North Korean attacks
against government sites and the New York Stock Exchange
demonstrate the distributed denial-of-service (DDoS) attack
(see http://www.foxnews.com/story/0,2933,530560,00.html and
http://www.foxnews.com/story/0,2933,530762,00.html).
I
have two concerns. The first is that most organizations do
not perform regular web application audits. We believe that
this is an essential part of a security or IT audit regimen.
Our Web Application Security Assessments generally reveal
that there are significant security flaws in web-based applications.
These flaws must be detected and remediated before your organization
becomes a victim of organized criminals seeking to earn a
living
off of companies with shoddy controls. I need not mention
the everyday hacker and the risks they pose. These are already
well known.
I am also concerned that
news articles represent that there is little that can be done
to respond to a denial-of-service
attack. That is not only incorrect, but it leads to an acceptance
of a serious risk without proper evaluation. Cisco wrote
a fantastic white paper on this issue (see http://www.cisco.com/en/US/tech/tk59/technologies_white_paper09186a0080174a5b.shtml#prevention).
I
use three slides in my IT Audit and Security Boot Camp to
show the before, during and after examples of a DDoS. One of
these slides is below. It shows how to successfully deal
with
a DDoS:
In the example shown,
the ISP for the client is using a cleansing agent to identify
the DDoS messages and filter them out. This cleansing agent
is a massive software application that is invoked when a DDoS
is detected for supported clients. We suggest that you check
with your ISP to determine if they have this type of software
and what the cost of invoking the product is. If they do not
have a product, then it is time to look for a new ISP. Do not
wait until the bot army of compromised machines is pointed
at your site and commences firing. The bad guys are constantly
scouring the Internet and phishing in their quest for poorly
secured machines. Once found, they
take control of the machines, place their software on them,
and organize them into a battle group for hire. They then
rent out their bot armies to those who can pay. In some cases,
even
governments have been known to use bot armies. The Russians
apparently did prior to and during their attack on Georgia
on August 7, 2007. It crippled the ability of Georgia to
respond to the attack. On July 4, North Korea may have done
the same
thing to our government.
The lesson here is do
not let complacency or a feeling of hopelessness prevent your
firm from building
the cyber defenses
needed in today’s environment. You need to be able to
ensure controls are in place in your network and the networks
that connect to yours. You also need to be able to respond
to a cyber attack and deflect a DDoS by a massive bot army.
All of these things can be achieved. Given the risks of the
last few weeks, now is the time to start a detailed risk analysis
of the client-side and application risks.
The opinions expressed
in this article are mine and mine alone. I look forward
to receiving your comments on this article and answering
any questions
you may have. You can email me at Gordon@canaudit.com.
If you
would like to receive articles like this in the future
directly, please opt-in to our distribution list at http://www.canaudit.com/mailinglist.html.
As
always, I like to provide an incentive for my clients to
conduct the required audit work. I believe that the web application
and external risk is very high. Therefore, I am willing to
price an Internet Review and Web Application Security Assessment
for our clients at $12,500. To qualify, the audit must be
confirmed
by August 31, 2009 and performed in August or September of
2009. If purchased separately, the cost is $23,000. For those
clients who want to review both the internal and external
security risks, we will provide a 10% discount on our IT Security
Baseline
or Network Penetration Audit. To qualify, the audit must
be confirmed by August 31, 2009 and performed in September
or
October of 2009.
If you would like
additional information on how Canaudit can assist you with
risk identification
and remediation
methodologies,
please contact Tamra Savage Jones at (805) 583-3723 or by
emailing her at Tamra@canaudit.com.
Canaudit specializes in a variety of information
system and technology audits, ranging from periodic network
penetration testing to full network and operating system security
review. Our tailored audits provide an objective, disciplined,
and in-depth analysis to evaluate and improve the effectiveness
of risk management, control and security within your organization’s
technological environment.
For interest in Canaudit
to perform an IT audit for your organization, please email Gordon@canaudit.com or Tamra@canaudit.com,
or call (805) 583-3723.
Canaudit provides quality seminars to
various organizations including audit and security chapters
and major corporations. These seminars range from technical
information system audit classes to internal audit classes
aimed at everyone from an introductory level up to management.
With nearly 20 courses to choose from, we are sure to have
one that will meet your individual needs. In addition to chapter
and private seminars, Canaudit also holds public courses. For
more information and to register for upcoming public courses,
visit http://www.canaudit.com/seminars.html.
For additional information
or to schedule a Canaudit seminar, please email Brenna@canaudit.com
or call (805) 583-3723.
|