Chapter 13 Information Security: Barbarians at the Gateway (and Just About Everywhere Else)

Bad Apples

While some of the more sensational exploits involve criminal gangs, research firm Gartner estimates that 70 percent of loss-causing security incidents involve insiders.J. Mardesich, “Ensuring the Security of Stored Data,” CIO Strategy Center, 2009. Rogue employees can steal secrets, install malware, or hold a firm hostage. Check processing firm Fidelity National Information Services was betrayed when one of its database administrators lifted personal records on 2.3 million of the firm’s customers and illegally sold them to direct marketers.

And it’s not just firm employees. Many firms hire temporary staffers, contract employees, or outsource key components of their infrastructure. Other firms have been compromised by members of their cleaning or security staff. A contract employee working at Sentry Insurance stole information on 110,000 of the firm’s clients.J. Vijayan, “Software Consultant Who Stole Data on 110,000 People Gets Five-Year Sentence,” Computerworld, July 10, 2007.

Social Engineering

As P. T. Barnum is reported to have said, “There’s a sucker born every minute.” Con games that trick employees into revealing information or performing other tasks that compromise a firm are known as social engineering in security circles. In some ways, crooks have never had easier access to background information that might be used to craft a scam. It’s likely that a directory of a firm’s employees, their titles, and other personal details is online right now via social networks like LinkedIn and Facebook. With just a few moments of searching, a skilled con artist can piece together a convincing and compelling story.

Data aggregator ChoicePoint sold private information to criminals who posed as legitimate clients, compromising the names, addresses, and Social Security numbers of some 145,000 individuals. In this breach, not a single computer was compromised. Employees were simply duped into turning data over to crooks. Gaffes like that can be painful. ChoicePoint paid $15 million in a settlement with the Federal Trade Commission, suffered customer loss, and ended up abandoning once lucrative businesses.G. Anthes, “The Grill: Security Guru Ira Winkler Takes the Hot Seat,” Computerworld, July 28, 2008.

Phishing

PhishingA con executed using technology, typically targeted at acquiring sensitive information or tricking someone into installing malicious software. refers to cons executed through technology. The goal of phishing is to leverage the reputation of a trusted firm or friend to trick the victim into performing an action or revealing information. The cons are crafty. Many have masqueraded as a security alert from a bank or e-commerce site (“Our Web site has been compromised, click to log in and reset your password.”), a message from an employer, or even a notice from the government (“Click here to update needed information to receive your tax refund transfer.”). Sophisticated con artists will lift logos, mimic standard layouts, and copy official language from legitimate Web sites or prior e-mails. Gartner estimates that these sorts phishing attacks cost consumers $3.2 billion in 2007.L. Avivah, “Phishing Attacks Escalate, Morph, and Cause Considerable Damage,” Gartner, December 12, 2007.

Other phishing attempts might dupe a user into unwittingly downloading dangerous software (malware) that can do things like record passwords and keystrokes, provide hackers with deeper access to your corporate network, or enlist your PC as part of a botnet. One attempt masqueraded as a message from a Facebook friend, inviting the recipient to view a video. Victims clicking the link were then told they need to install an updated version of the Adobe Flash plug-in to view the clip. The plug in was really a malware program that gave phishers control of the infected user’s computer.B. Krebs, “‘Koobface’ Worm Resurfaces on Facebook, MySpace,” Washington Post, March 2, 2009. Other attempts have populated P2P networks (peer-to-peer file distribution systems such as BitTorrent) with malware-installing files masquerading as video games or other software, movies, songs, and pornography.

So-called spear phishing attacks specifically target a given organization or group of users. In one incident, employees of a medical center received e-mails purportedly from the center itself, indicating that the recipient was being laid off and offering a link to job counseling resources. The link really offered a software payload that recorded and forwarded any keystrokes on the victim’s PC.C. Garretson, “Spam that Delivers a Pink Slip,” NetworkWorld, November 1, 2006. And with this type of phishing, the more you know about a user, the more convincing it is to con them. Phishers using pilfered résumé information from Monster.com crafted targeted and personalized e-mails. The request, seemingly from the job site, advised users to download the “Monster Job Seeker Tool”; this “tool” installed malware that encrypted files on the victim’s PC, leaving a ransom note demanding payment to liberate a victim’s hard disk.T. Wilson, “Trojan On Monster.com Steals Personal Data,” Forbes, August 20, 2007.

Passwords

Many valuable assets are kept secure via just one thin layer of protection—the password. And if you’re like most users, your password system is a mess.F. Manjoo, “Fix Your Terrible, Insecure Passwords in Five Minutes,” Slate, November 12, 2009. With so many destinations asking for passwords, chances are you’re using the same password (or easily guessed variants) in a way that means getting just one “key” would open many “doors.” The typical Web user has 6.5 passwords, each of which is used at four sites, on average.N. Summers, “Building a Better Password,” Newsweek, October 19, 2009. Some sites force users to change passwords regularly, but this often results in insecure compromises. Users make only minor tweaks (e.g., appending the month or year); they write passwords down (in an unlocked drawer or Post-it note attached to the monitor); or they save passwords in personal e-mail accounts or on unencrypted hard drives.

The challenge questions offered by many sites to automate password distribution and reset are often pitifully insecure. What’s your mother’s maiden name? What elementary school did you attend? Where were you born? All are pretty easy to guess. One IEEE study found acquaintances could correctly answer colleagues’ secret questions 28 percent of the time, and those who did not know the person still guessed right at a rate of 17 percent. Plus, within three to six months, 16 percent of study participants forgot answers to their own security questions.R. Lemos, “Are Your ‘Secret Questions’ Too Easily Answered?” Technology Review, May 18, 2009. In many cases, answers to these questions can be easily uncovered online. Chances are, if you’ve got an account at a site like Ancestry.com, classmates.com, or Facebook, then some of your secret answers have already been exposed—by you! A Tennessee teen hacked into Sarah Palin’s personal Yahoo! account (gov.palin@yahoo.com) in part by correctly guessing where she met her husband. A similar attack hit staffers at Twitter, resulting in the theft of hundreds of internal documents, including strategy memos, e-mails, and financial forecasts, many of which ended up embarrassingly posted online.N. Summers, “Building a Better Password,” Newsweek, October 19, 2009.

Related to the password problem are issues with system setup and configuration. Many vendors sell software with a common default password. For example, for years, leading database products came with the default account and password combination “scott/tiger.” Any firm not changing default accounts and passwords risks having an open door. Other firms are left vulnerable if users set systems for open access—say turning on file sharing permission for their PC. Programmers, take note: well-designed products come with secure default settings, require users to reset passwords at setup, and also offer strong warnings when security settings are made weaker. But unfortunately, there are a lot of legacy products out there, and not all vendors have the insight to design for out-of-the-box security.

Malware

Any accessible computing device is a potential target for infiltration by malware. Malware (for malicious software) seeks to compromise a computing system without permission. Client PCs and a firm’s servers are primary targets, but as computing has spread, malware now threatens nearly any connected system running software, including mobile phones, embedded devices, and a firm’s networking equipment.

Some hackers will try to sneak malware onto a system via techniques like phishing. In another high-profile hacking example, infected USB drives were purposely left lying around government offices. Those seemingly abandoned office supplies really contained code that attempted to infiltrate government PCs when inserted by unwitting employees.

Machines are constantly under attack. Microsoft’s Internet Safety Enforcement Team claims that the mean time to infection for an unprotected PC is less than five minutes.J. Markoff, “A Robot Network Seeks to Enlist Your Computer,” New York Times, October 20, 2008. Oftentimes malware attempts to compromise weaknesses in software—either bugs, poor design, or poor configuration.

Years ago, most attacks centered on weaknesses in the operating system, but now malware exploits have expanded to other targets, including browsers, plug-ins, and scripting languages used by software. BusinessWeek reports that Adobe has replaced Microsoft as the primary means by which hackers try to infect or take control of PCs. Even trusted Web sites have become a conduit to deliver malware payloads. More than a dozen sites, including those of the New York Times, USA Today, and Nature, were compromised when seemingly honest advertising clients switched on fake ads that exploit Adobe software.A. Ricadela, “Can Adobe Beat Back the Hackers?” BusinessWeek, November 19, 2009. Some attacks were delivered through Flash animations that direct computers to sites that scan PCs, installing malware payloads through whatever vulnerabilities are discovered. Others circulated via e-mail through PDF triggered payloads deployed when a file was loaded via Acrobat Reader. Adobe is a particularly tempting target, as Flash and Acrobat Reader are now installed on nearly every PC, including Mac and Linux machines.

Malware goes by many names. Here are a few of the more common terms you’re likely to encounter.Portions adapted from G. Perera, “Your Guide to Understanding Malware,” LaptopLogic.com, May 17, 2009.

Methods of infection are as follows:

• Viruses. Programs that infect other software or files. They require an executable (a running program) to spread, attaching to other executables. Viruses can spread via operating systems, programs, or the boot sector or auto-run feature of media such as DVDs or USB drives. Some applications have executable languages (macros) that can also host viruses that run and spread when a file is open.
• Worms. Programs that take advantage of security vulnerability to automatically spread, but unlike viruses, worms do not require an executable. Some worms scan for and install themselves on vulnerable systems with stunning speed (in an extreme example, the SQL Slammer worm infected 90 percent of vulnerable software worldwide within just ten minutes).M. Broersma, “Slammer—the First ‘Warhol’ Worm?” CNET, February 3, 2003.
• Trojans. Exploits that, like the mythical Trojan horse, try to sneak in by masquerading as something they’re not. The payload is released when the user is duped into downloading and installing the malware cargo, oftentimes via phishing exploits.

While the terms above cover methods for infection, the terms below address the goal of the malware:

• Botnets or zombie networks. Hordes of surreptitiously infected computers linked and controlled remotely by a central command. Botnets are used in crimes where controlling many difficult-to-identify PCs is useful, such as when perpetrating click fraud, sending spam, registering accounts that use CAPTCHAsAn acronym for Completely Automated Public Turing Test to Tell Computers and Humans Apart. CAPTCHAs are those scrambled character images that many sites require to submit some sort of entry (account setup, ticket buying). CAPTCHAs were developed because computers have difficulty discerning letters that are distorted or mixed inside a jumbled graphic. CAPTCHAs are meant to be a Turing Test—a test to distinguish if a task is being performed by a computer or a human.G. Keizer, “Botnet Busts Newest Hotmail CAPTCHA,” Computerworld, February 19, 2009. (those scrambled character images meant to thwart things like automated account setup or ticket buying), executing “dictionary” password cracking attempts, or launching denial-of-service attacks.
• Malicious adware. Programs installed without full user consent or knowledge that later serve unwanted advertisements.
• Spyware. Software that surreptitiously monitors user actions, network traffic, or scans for files.
• Keylogger. Type of spyware that records user keystrokes. Keyloggers can be either software-based or hardware, such as a recording “dongle” that is plugged in between a keyboard and a PC.
• Screen capture. Variant of the keylogger approach. This category of software records the pixels that appear on a user’s screen for later playback in hopes of identifying proprietary information.
• Blended threats. Attacks combining multiple malware or hacking exploits.

Compromising Web Sites

Some exploits directly target poorly designed and programmed Web sites. Consider the SQL injection technique. It zeros in on a sloppy programming practice where software developers don’t validate user input.

It works like this. Imagine that you visit a Web site and are asked to enter your user ID in a field on a Web page (say your user ID is smith). A Web site may be programmed to take the data you enter from the Web page’s user ID field (smith), then add it to a database command (creating the equivalent of a command that says “find the account for ‘smith’”). The database then executes that command.

But Web sites that don’t verify user entries and instead just blindly pass along entered data are vulnerable to attack. Hackers with just a rudimentary knowledge of SQL could type actual code fragments into the user ID field, appending this code to statements executed by the site (see sidebar for a more detailed description). Such modified instructions could instruct the Web site’s database software to drop (delete) tables, insert additional data, return all records in a database, or even redirect users to another Web site that will scan clients for weaknesses, then launch further attacks. Security expert Ben Schneier noted a particularly ghastly SQL injection vulnerability in the publicly facing database for the Oklahoma Department of Corrections, where “anyone with basic SQL knowledge could have registered anyone he wanted as a sex offender.”B. Schneier, “Oklahoma Data Leak,” Schneier on Security, April 18, 2008.

Not trusting user input is a cardinal rule of programming, and most well-trained programmers know to validate user input. But there’s a lot of sloppy code out there, which hackers are all too eager to exploit. IBM identifies SQL injection as the fastest growing security threat, with over half a million attack attempts recorded each day.A. Wittmann, “The Fastest-Growing Security Threat,” InformationWeek, November 9, 2009. Some vulnerable systems started life as quickly developed proofs of concepts, and programmers never went back to add the needed code to validate input and block these exploits. Other Web sites may have been designed by poorly trained developers who have moved on to other projects, by staff that have since left the firm, or where development was outsourced to another firm. As such, many firms don’t even know if they suffer from this vulnerability.

SQL injection and other application weaknesses are particularly problematic because there’s not a commercial software patch or easily deployed piece of security software that can protect a firm. Instead, firms have to meticulously examine the integrity of their Web sites to see if they are vulnerable.While some tools exist to automate testing, this is by no means as easy a fix as installing a commercial software patch or virus protection software.

Related programming exploits go by names such as cross-site scripting attacks and HTTP header injection. We’ll spare you the technical details, but what this means for both the manager and the programmer is that all systems must be designed and tested with security in mind. This includes testing new applications, existing and legacy applications, partner offerings, and SaaS (software as a service) applications—everything. Visa and MasterCard are among the firms requiring partners to rigorously apply testing standards. Firms that aren’t testing their applications will find they’re locked out of business; if caught with unacceptable breaches, such firms may be forced to pay big fines and absorb any costs associated with their weak practices.“Information Security: Why Cybercriminals Are Smiling,” Knowledge@Wharton, August 19, 2009.

Network Threats

The network itself may also be a source of compromise. Recall that the TJX hack happened when a Wi-Fi access point was left open and undetected. A hacker just drove up and performed the digital equivalent of crawling through an open window. The problem is made more challenging since wireless access points are so inexpensive and easy to install. For less than $100, a user (well intentioned or not) could plug in to an access point that could provide entry for anyone. If a firm doesn’t regularly monitor its premises, its network, and its network traffic, it may fall victim.

Other troubling exploits have targeted the very underpinning of the Internet itself. This is the case with so-called DNS cache poisoning. The DNS, or domain name service, is a collection of software that maps an Internet address, such as (http://www.bc.edu), to an IP address, such as 136.167.2.220. 220 (see Chapter 12 "A Manager’s Guide to the Internet and Telecommunications" for more detail). DNS cache poisoning exploits can redirect this mapping and the consequences are huge. Imagine thinking that you’re visiting your bank’s Web site, but instead your network’s DNS server has been poisoned so that you really visit a carefully crafted replica that hackers use to steal your log-in credentials and drain your bank account. A DNS cache poisoning attack launched against one of China’s largest ISPs redirected users to sites that launched malware exploits, targeting weaknesses in RealPlayer, Adobe Flash, and Microsoft’s ActiveX technology, commonly used in browsers.J. London, “China Netcom Falls Prey to DNS Cache Poisoning,” Computerworld, August 22, 2008.

Physical Threats

A firm doesn’t just have to watch out for insiders or compromised software and hardware; a host of other physical threats can grease the skids to fraud, theft, and damage. Most large firms have disaster-recovery plans in place. These often include provisions to backup systems and data to off-site locales, to protect operations and provide a fall back in the case of disaster. Such plans increasingly take into account the potential impact of physical security threats such as terrorism, or vandalism, as well.

Anything valuable that reaches the trash in a recoverable state is also a potential security breach. Hackers and spies sometimes practice dumpster divingCombing through trash to identify valuable assets., sifting through trash in an effort to uncover valuable data or insights that can be stolen or used to launch a security attack. This might include hunting for discarded passwords written on Post-it notes, recovering unshredded printed user account listings, scanning e-mails or program printouts for system clues, recovering tape backups, resurrecting files from discarded hard drives, and more.

Other compromises might take place via shoulder surfingGaining compromising information through observation (as in looking over someone’s shoulder)., simply looking over someone’s shoulder to glean a password or see other proprietary information that might be displayed on a worker’s screen.

Firms might also fall victim to various forms of eavesdropping, such as efforts to listen into or record conversations, transmissions, or keystrokes. A device hidden inside a package might sit inside a mailroom or a worker’s physical inbox, scanning for open wireless connections, or recording and forwarding conversations.J. Robertson, “Hackers Mull Physical Attacks on a Networked World,” San Francisco Chronicle, August 8, 2008. Other forms of eavesdropping can be accomplished via compromised wireless or other network connections, malware keylogger or screen capture programs, as well as hardware devices such as replacement keyboards with keyloggers embedded inside, microphones to capture the slightly unique and identifiable sound of each key being pressed, programs that turn on built-in microphone or cameras that are now standard on many PCs, or even James Bond-style devices using Van Eck techniques that attempt to read monitors from afar by detecting their electromagnetic emissions.

Frameworks, Standards, and Compliance

Developing organizational security is a daunting task. You’re in an arms race with adversaries that are tenacious and constantly on the lookout for new exploits. Fortunately, no firm is starting from scratch—others have gone before you and many have worked together to create published best practices.

There are several frameworks, but perhaps the best known of these efforts comes from the International Organization for Standards (ISO), and is broadly referred to as ISO27k or the ISO 27000 series. According to ISO.org, this evolving set of standards provides “a model for establishing, implementing, operating, monitoring, reviewing, maintaining, and improving an Information Security Management System.”

Firms may also face compliance requirements—legal or professionally binding steps that must be taken. Failure to do so could result in fine, sanction, and other punitive measures. At the federal level, examples include HIPAA (the Health Insurance Portability and Accountability Act), which regulates health data; the Graham-Leach-Bliley Act, which regulates financial data; and the Children’s Online Privacy Protection Act, which regulates data collection on minors. U.S. government agencies must also comply with FISMA (the Federal Information Security Management Act), and there are several initiatives at the other government levels. By 2009, some level of state data breach laws had been passed by over thirty states, while multinationals face a growing number of statues throughout the world. Your legal team and trade associations can help you understand your domestic and international obligations. Fortunately, there are often frameworks and guidelines to assist in compliance. For example, the ISO standards include subsets targeted at the telecommunications and health care industries, and major credit card firms have created the PCI (payment card industry) standards. And there are skilled consulting professionals who can help bring firms up to speed in these areas, and help expand their organizational radar as new issues develop.

Here is a word of warning on frameworks and standards: compliance does not equal security. Outsourcing portions security efforts without a complete, organizational commitment to being secure can also be dangerous. Some organizations simply approach compliance as a necessary evil: a sort of checklist that can reduce the likelihood of a lawsuit or other punitive measure.M. Davis, “What Will It Take?” InformationWeek, November 23, 2009. While you want to make sure you’re doing everything in your power not to get sued, this isn’t the goal. The goal is taking all appropriate measures to ensure that your firm is secure for your customers, employees, shareholders, and others. Frameworks help shape your thinking and expose things you should do, but security doesn’t stop there—this is a constant, evolving process that needs to pervade the organization from the CEO suite and board, down to front line workers and potentially out to customers and partners. And be aware of the security issues associated with any mergers and acquisitions. Bringing in new firms, employees, technologies, and procedures means reassessing the security environment for all players involved.

Education, Audit, and Enforcement

Security is as much about people, process, and policy, as it is about technology.

From a people perspective, the security function requires multiple levels of expertise. Operations employees are involved in the day-to-day monitoring of existing systems. A group’s R&D function is involved in understanding emerging threats and reviewing, selecting, and implementing updated security techniques. A team must also work on broader governance issues. These efforts should include representatives from specialized security and broader technology and infrastructure functions. It should also include representatives from general counsel, audit, public relations, and human resources. What this means is that even if you’re a nontechnical staffer, you may be brought in to help a firm deal with security issues.

Processes and policies will include education and awareness—this is also everyone’s business. As the Vice President of Product Development at security firm Symantec puts it, “We do products really well, but the next step is education. We can’t keep the Internet safe with antivirus software alone.”D. Goldman, “Cybercrime: A Secret Underground Economy,” CNNMoney, September 17, 2009. Companies should approach information security as a part of their “collective corporate responsibility…regardless of whether regulation requires them to do so.”Knowledge@Wharton, “Information Security: Why Cybercriminals Are Smiling,” August 19, 2009.

For a lesson in how important education is, look no further than the head of the CIA. Former U.S. Director of Intelligence John Deutch engaged in shockingly loose behavior with digital secrets, including keeping a daily journal of classified information—some 1,000+ pages—on memory cards he’d transport in his shirt pocket. He also downloaded and stored Pentagon information, including details of covert operations, at home on computers that his family used for routine Internet access.N. Lewis, “Investigation Of Ex-Chief Of the C.I.A. Is Broadened,” New York Times, September 17, 2000.

Employees need to know a firm’s policies, be regularly trained, and understand that they will face strict penalties if they fail to meet their obligations. Policies without eyes (audit) and teeth (enforcement) won’t be taken seriously. Audits include real-time monitoring of usage (e.g., who’s accessing what, from where, how, and why; sound the alarm if an anomaly is detected), announced audits, and surprise spot checks. This function might also stage white hat demonstration attacks—attempts to hunt for and expose weaknesses, hopefully before hackers find them. Frameworks offer guidelines on auditing, but a recent survey found most organizations don’t document enforcement procedures in their information security policies, that more than one-third do not audit or monitor user compliance with security policies, and that only 48 percent annually measure and review the effectiveness of security policies.A. Matwyshyn, Harboring Data: Information Security, Law, and The Corporation (Palo Alto, CA: Stanford University Press, 2009).

A firm’s technology development and deployment processes must also integrate with the security team to ensure that from the start, applications, databases, and other systems are implemented with security in mind. The team will have specialized skills and monitor the latest threats and are able to advise on precautions necessary to be sure systems aren’t compromised during installation, development, testing, and deployment.

What Needs to Be Protected and How Much Is Enough?

A worldwide study by PricewaterhouseCoopers and Chief Security Officer magazine revealed that most firms don’t even know what they need to protect. Only 33 percent of executives responded that their organizations kept accurate inventory of the locations and jurisdictions where data was stored, and only 24 percent kept inventory of all third parties using their customer data.A. Matwyshyn, Harboring Data: Information Security, Law, and The Corporation (Palo Alto, CA: Stanford University Press, 2009). What this means is that most firms don’t even have an accurate read on where their valuables are kept, let alone how to protect them.

So information security should start with an inventory-style auditing and risk assessment. Technologies map back to specific business risks. What do we need to protect? What are we afraid might happen? And how do we protect it? Security is an economic problem, involving attack likelihood, costs, and prevention benefits. These are complex trade-offs that must consider losses from theft or resources, systems damage, data loss, disclosure of proprietary information, recovery, downtime, stock price declines, legal fees, government and compliance penalties, and intangibles such as damaged firm reputation, loss of customer and partner confidence, industry damage, promotion of adversary, and encouragement of future attacks.

While many firms skimp on security, firms also don’t want to misspend, targeting exploits that aren’t likely, while underinvesting in easily prevented methods to thwart common infiltration techniques. Hacker conventions like DefCon can show some really wild exploits. But it’s up to the firm to assess how vulnerable it is to these various risks. The local donut shop has far different needs than a military installation, law enforcement agency, financial institution, or firm housing other high-value electronic assets. A skilled risk assessment team will consider these vulnerabilities and what sort of countermeasure investments should take place.

Economic decisions usually drive hacker behavior, too. While in some cases attacks are based on vendetta or personal reasons, in most cases exploit economics largely boils down to

Adversary ROI = Asset value to adversary – Adversary cost.

An adversary’s costs include not only the resources, knowledge, and technology required for the exploit, but also the risk of getting caught. Make things tough to get at, and lobbying for legislation that imposes severe penalties on crooks can help raise adversary costs and lower your likelihood of becoming a victim.

Technology’s Role

Technical solutions often involve industrial strength variants of the previously discussed issues individuals can employ, so your awareness is already high. Additionally, an organization’s approach will often leverage multiple layers of protection and incorporate a wide variety of protective measures.

Patch. Firms must be especially vigilant to pay attention to security bulletins and install software updates that plug existing holes, (often referred to as patches). Firms that don’t plug known problems will be vulnerable to trivial and automated attacks. Unfortunately, many firms aren’t updating all components of their systems with consistent attention. With operating systems automating security update installations, hackers have moved on to application targets. But a major study recently found that organizations took at least twice as long to patch application vulnerabilities as they take to patch operating system holes.S. Wildstrom, “Massive Study of Net Vulnerabilities: They’re Not Where You Think They Are,” BusinessWeek, September 14, 2009. And remember, software isn’t limited to conventional PCs and servers. Embedded systems abound, and connected, yet unpatched devices are vulnerable. Malware has infected everything from unprotected ATM machinesP. Lilly, “Hackers Targeting Windows XP-Based ATM Machines,” Maximum PC, June 4, 2009. to restaurant point-of-sale systemsR. McMillan, “Restaurants Sue Vendors after Point-of-Sale Hack,” CIO, December 1, 2009. to fighter plane navigation systems.C. Matyszczyk, “French Planes Grounded by Windows Worm,” CNET, February 8, 2009.

As an example of unpatched vulnerabilities, consider the DNS cache poisoning exploit described earlier in this chapter. The discovery of this weakness was one of the biggest security stories the year it was discovered, and security experts saw this as a major threat. Teams of programmers worldwide raced to provide fixes for the most widely used versions of DNS software. Yet several months after patches were available, roughly one quarter of all DNS servers were still unpatched and exposed.IBM, X-Force Threat Report: 2008 Year in Review, January 2009.

To be fair, not all firms delay patches out of negligence. Some organizations have legitimate concerns about testing whether the patch will break their system or whether the new technology contains a change that will cause problems down the road.For example, the DNS security patch mentioned was incompatible with the firewall software deployed at some firms. And there have been cases where patches themselves have caused problems. Finally, many software updates require that systems be taken down. Firms may have uptime requirements that make immediate patching difficult. But ultimately, unpatched systems are an open door for infiltration.

Lock down hardware. Firms range widely in the security regimes used to govern purchase through disposal system use. While some large firms such as Kraft are allowing employees to select their own hardware (Mac or PC, desktop or notebook, iPhone or BlackBerry),N. Wingfield, “It’s a Free Country…So Why Can’t I Pick the Technology I Use in the Office?” Wall Street Journal, November 15, 2009. others issue standard systems that prevent all unapproved software installation and force file saving to hardened, backed-up, scanned, and monitored servers. Firms in especially sensitive industries such as financial services may regularly reimage the hard drive of end-user PCs, completely replacing all the bits on a user’s hard drive with a pristine, current version—effectively wiping out malware that might have previously sneaked onto a user’s PC. Other lock-down methods might disable the boot capability of removable media (a common method for spreading viruses via inserted discs or USBs), prevent Wi-Fi use or require VPN encryption before allowing any network transmissions, and more. The cloud helps here, too. (See Chapter 10 "Software in Flux: Partly Cloudy and Sometimes Free".) Employers can also require workers to run all of their corporate applications inside a remote desktop where the actual executing hardware and software is elsewhere (likely hosted as a virtual machine session on the organization’s servers), and the user is simply served an image of what is executing remotely. This seals the virtual PC off in a way that can be thoroughly monitored, updated, backed up, and locked down by the firm.

In the case of Kraft, executives worried that the firm’s previously restrictive technology policies prevented employees from staying in step with trends. Employees opting into the system must sign an agreement promising they’ll follow mandated security procedures. Still, financial services firms, law offices, health care providers, and others may need to maintain stricter control, for legal and industry compliance reasons.

Lock down the network. Network monitoring is a critical part of security, and a host of technical tools can help.

Firms employ firewallsA system that acts as a control for network traffic, blocking unauthorized traffic while permitting acceptable use. to examine traffic as it enters and leaves the network, potentially blocking certain types of access, while permitting approved communication. Intrusion detection systemsA system that monitors network use for potential hacking attempts. Such a system may take preventative action to block, isolate, or identify attempted infiltration, and raise further alarms to warn security personnel. specifically look for unauthorized behavior, sounding the alarm and potentially taking action if something seems amiss. Some firms deploy honeypotsA seemingly tempting, but bogus target meant to draw hacking attempts. By monitoring infiltration attempts against a honeypot, organizations may gain insight into the identity of hackers and their techniques, and they can share this with partners and law enforcement.—bogus offerings meant to distract attackers. If attackers take honeypot bait, firms may gain an opportunity to recognize the hacker’s exploits, identify the IP address of intrusion, and take action to block further attacks and alert authorities.

Many firms also deploy blacklistsPrograms that deny the entry or exit of specific IP addresses, products, Internet domains, and other communication restrictions.—denying the entry or exit of specific IP addresses, products, Internet domains, and other communication restrictions. While blacklists block known bad guys, whitelistsHighly restrictive programs that permit communication only with approved entities and/or in an approved manner. are even more restrictive—permitting communication only with approved entities or in an approved manner.

These technologies can be applied to network technology, specific applications, screening for certain kinds of apps, malware signatures, and hunting for anomalous patterns. The latter is important, as recent malware has become polymorphic, meaning different versions are created and deployed in a way that their signature, a sort of electronic fingerprint often used to recognize malicious code, is slightly altered. This also helps with zero-day exploits, and in situations where whitelisted Web sites themselves become compromised.

Many technical solutions, ranging from network monitoring and response to e-mail screening, are migrating to “the cloud.” This can be a good thing—if network monitoring software immediately shares news of a certain type of attack, defenses might be pushed out to all clients of a firm (the more users, the “smarter” the system can potentially become—again we see the power of network effects in action).

Lock down partners. Insist partner firms are compliant, and audit them to ensure this is the case. This includes technology providers and contract firms, as well as value chain participants such as suppliers and distributors. Anyone who touches your network is a potential point of weakness. Many firms will build security expectations and commitments into performance guarantees known as service level agreements (SLAs).

Lock down systems. Audit for SQL injection and other application exploits. The security team must constantly scan exploits and then probe its systems to see if it’s susceptible, advising and enforcing action if problems are uncovered. This kind of auditing should occur with all of a firm’s partners.

Access controls can also compartmentalize data access on a need-to-know basis. Such tools can not only enforce access privileges, they can help create and monitor audit trails to help verify that systems are not being accessed by the unauthorized, or in suspicious ways.

Audit trails are used for deterring, identifying, and investigating these cases. Recording, monitoring, and auditing access allows firms to hunt for patterns of abuse. Logs can detail who, when, and from where assets are accessed. Giveaways of nefarious activity may include access from unfamiliar IP addresses, from nonstandard times, accesses that occur at higher than usual volumes, and so on. Automated alerts can put an account on hold or call in a response team for further observation of the anomaly.

Single-sign-on tools can help firms offer employees one very strong password that works across applications, is changed frequently (or managed via hardware cards or mobile phone log-in), and can be altered by password management staff.

Multiple administrators should jointly control key systems. Major configuration changes might require approval of multiple staffers, as well as the automatic notification of concerned personnel. And firms should employ a recovery mechanism to regain control in the event that key administrators are incapacitated or uncooperative. This balances security needs with an ability to respond in the event of a crisis. Such a system was not in place in the earlier described case of the rogue IT staffer who held the city of San Francisco’s networks hostage by refusing to give up vital passwords.

Have failure and recovery plans. While firms work to prevent infiltration attempts, they should also have provisions in place that plan for the worst. If a compromise has taken place, what needs to be done? Do stolen assets need to be devalued (e.g., accounts terminated, new accounts issued)? What should be done to notify customers and partners, educate them, and advise them through any necessary responses? Who should work with law enforcement and with the media? Do off-site backups or redundant systems need to be activated? Can systems be reliably restored without risking further damage?

Best practices are beginning to emerge. While postevent triage is beyond the scope of our introduction, the good news is that firms are now sharing data on breaches. Given the potential negative consequences of a breach, organizations once rarely admitted they’d been compromised. But now many are obligated to do so. And the broad awareness of infiltration both reduces organizational stigma in coming forward, and allows firms and technology providers to share knowledge on the techniques used by cybercrooks.

Information security is a complex, continually changing, and vitally important domain. The exploits covered in this chapter seem daunting, and new exploits constantly emerge. But your thinking on key issues should now be broader. Hopefully you’ve now embedded security thinking in your managerial DNA, and you are better prepared to be a savvy system user and a proactive participant working for your firm’s security. Stay safe!