This chapter is from the book
This chapter is from the book
This chapter is from the book
10.2 Campaign-Targeted Phishing
Phishing has without a doubt become one of the most widespread risks affecting Internet users today. When we look at phishing and the role that it may play in an election campaign, we can readily envision several incremental risks that present themselves beyond the traditional theft of confidential information.
10.2.1 Profit-Motivated Phishing
Profit-motivated, event-based phishing is certainly not new. It has been seen in the past on numerous occasions leading up to and following significant events worldwide. For example, this type of attack was seen after natural disasters such as the Indian Ocean tsunami in 2004 [66] and Hurricane Katrina in 2005 [220, 251]. It was also seen in conjunction with sporting events, such as the 2006 and 2010 FIFA World Cup [275].
Election-related phishing has been observed in the past. During the 2004 federal election, phishers targeted the Kerry–Edwards campaign [370], a campaign that was acknowledged as being at the forefront of leveraging the Internet for communications. At least two distinct types of phishing were observed. In one case, phishers set up a fictitious web site to solicit online campaign contributions shortly after the Democratic National Convention; this site stole the victim's credit card number, among other information. In the second case, phishers asked recipients to call a for-fee 1-900 number, for which the victim would subsequently be charged $1.99 per minute [417]. This is a prime example of how such attacks can cross technology boundaries to appear even more convincing. The perpetrators of these two attacks were never caught.
When considering the 2004 election as a whole, phishing presented only a marginal risk. At the time, phishing was still in its infancy, and had yet to grow into the epidemic that can be observed today. When assessing the potential risk of phishing in conjuction with the 2008 federal election, however, we find ourselves in a much different position. Candidates have flocked to the Internet, seeing it as a key means to communicate with constituents and to raise campaign contributions.
We performed an analysis of campaign web sites in an attempt to determine to what degree they allow contributions to be made online. We discovered that every candidate provided a mechanism by which supporters could make a donation online. All of the web sites on which contributions could be made leveraged SSL as a means to secure the transaction. We also noted the domain of each contribution site. In numerous cases, would-be contributors were redirected to a third-party site, which sat on a different primary domain. Table 10.3 lists both the original domain, and the web site to which the user is redirected.
Table 10.3. An analysis of 2008 federal candidate web sites and the sites to which contributors are directed to. The sites to which contributors are redirected are legitimate, but the fact that they are often different from the original site increases the risk for confusion and thereby the risk that a phishing attack with a similar design would succeed.
|
Domain Name |
Redirects to |
|
barackobama.com |
|
|
brownback.com |
|
|
chrisdodd.com |
|
|
cox2008.com |
|
|
mikehuckabee.com |
|
|
gilmoreforpresident.com |
|
|
gohunter08.com |
|
|
hillaryclinton.com |
|
|
joebiden.com |
|
|
johnedwards.com |
|
|
johnmccain.com |
|
|
joinrudy2008.com |
|
|
mittromney.com |
|
|
richardsonforpresident.com |
|
|
ronpaul2008.com |
|
|
teamtancredo.com |
|
|
tommy2008.com |
This redirection was the result of third-party consulting, media, and online advocacy firms being used to assist in the running of the campaign, including the processing of online campaign contributions. This practice does not present a security risk in and of itself, nor is it an indication that phishing is taking place; however, the change in the top-level domain may add to the confusion of potential contributors, who tend to err on the side of caution. It also indicates that additional parties may be involved in the gathering and processing of personal information on behalf of a campaign, increasing the overall exposure of the credit card numbers processed during fundraising.
It should also be noted that the redirection used here is not necessary, and that the contribution site could just as easily remain in the same top-level domain, as a subdomain hosted by the third party for processing. To do so simply requires the appropriate configuration of the primary domain's DNS records. In fact, the majority of the remaining candidates have chosen to follow this path. Future research may also reveal whether those donation sites that do live under the campaign's domain name are, in fact, hosted on the same physical network as the campaign web site or on another third-party payment processor's network.
Figure 10.6 provides a sample of the information collected during an online contribution. We found that forms were fairly consistent in the type of information that was collected, while (not surprisingly) varying from a visual perspective.
)
Figure 10.6 A sample form from one candidate's web site allowing visitors to make contributions online. This is a legitimate site. Given that typical Internet users would not be well acquainted with the domains associated with political candidates, there is a risk that phishers might use a similarly designed web site to collect credentials from unsuspecting victims.
The ability to process credit card transactions on an authentic campaign web site may provide an unexpected benefit to online identity thieves. One tactic regularly employed by those peddling in stolen credit cards is to process a very small transaction so as to validate a credit card as legitimate [48]. Thieves began using this technique in early 2007 on online charity web sites, but it has long been used on other types of online payment sites. Such a small transaction is unlikely to be noticed by the credit card holder and is unlikely to be flagged by the party processing the transaction.
Of course, not all contributions would necessarily be helpful. Attackers might seek to disrupt a candidate's fundraising efforts by initiating illegitimate payments to create confusion. If performed en masse, the widespread contribution of small, random amounts of money, from thousands or tens of thousands of stolen credit cards, would certainly have a negative effect. While there is a slight chance such an attack might remain stealth, it is more likely that it will be noticed, making it nearly impossible to differentiate legitimate contributions from fraudulent donations. Thus a significant burden would be placed on the affected candidates by diluting legitimate contributions with those that were not initiated by the credit card owners.
The increased collection of online campaign contributions also provides a ripe opportunity for phishers to target members of the unsuspecting public. Candidates and their parties regularly communicate with voters through email, as demonstrated in Figure 10.7. Phishing involves the use of email to lure a victim to a fictitious web site that attempts to steal confidential information from the victim [91]. While it is unreasonable to expect campaigns not to solicit contributions using email as a medium, they would be well advised to follow best practices that have been set by other online entities heavily prone to phishing. (A number of excellent resources are available through the Anti-Phishing Working Group [313], including a report funded by the U.S. Department of Homeland Security [101] that discusses the problem in depth and suggests best practices for organizations to communicate safely with their constituents.) However, whether or not the candidate uses email for contribution requests, a phisher may pose as a candidate and ask the recipients of his or her email for money. The typical goal would be to steal the credentials of the victims.
)
Figure 10.7 A portion of a legitimate fundraising email, which allows the recipient to click on the hyperlinked "Contribute" button to support the campaign. This approach would be very easy for a phisher to mimic in an effort to make people submit their credentials to the phisher, thinking they are contributing. Of course, phishers can use inflammatory texts (even more so than political candidates) as calls for action. The authors of this book were able to register the domain democratic-party.us, which would be suitable in such an attack, and found a wealth of other cousin name domains available for both parties. Thus, whereas financial institutions typically have registered cousin name domains to defend against abuse, political parties and candidates have not.
Phishers can increase their success rate by registering domain names that are typos or cousin domains of their target, a tactic already discussed in some depth in this chapter. For example, a phisher targeting John Edwards might elect to register donatejohnedwards.com. Additionally, phishers may simply create subdomains for primary domains that they already own. A phisher who buys the domain donatefor2008.com, for example, might simply add DNS records for johnedwards.donatefor2008.com and ronpaul.donatefor2008.com, among others. These domain names could then be referenced in the phishing emails sent to potential victims. When clicked on, the link would drive the victim to the fictitious web site.
As we have observed, a significant number of typo domain names have already been registered, or are available to be registered, by parties who are acting in bad faith. Many of these domain names appear so similar to the legitimate domain name that the unsuspecting eye of a potential victim would not notice if directed to one of these sites. Campaigns can take clear and immediate steps to purchase typo domains prior to them falling into the wrong hands. As of this writing, few have done so.
More difficult, however, is the acquisition of cousin domain names. As discussed previously, a significant number of cousin domain names have been registered for both speculative and advertising purposes. Given the near-infinite number of possible cousin domain names, it is unlikely that a campaign could acquire all possibilities. This fact of life provides phishers with the opportunity to register a domain name that may appear similar to the legitimate campaign's web site.
Yet another type of attack might use a spoofed email that appears to come from a political party or candidate to entice recipients to open attachments, thereby infecting their machines with malicious code. Again, this may be done either with the direct goal of spreading malicious code or to deliver a below-the-belt blow to political candidates who rely heavily on the Internet for their communication with constituents.
Even without the registration of a similar domain name, phishers will undoubtedly continue to succeed in constructing emails and web sites that are obvious to detect by a trained eye, but perhaps not so obvious to those who continue to fall victim to them.