Security Means Making Trade-Offs to Manage Risks
Security isn’t having the strongest lock or the best anti-virus software — security is about making trade-offs to manage risk, something we do in many contexts throughout the day. When you consider crossing the street in the middle of the block rather than at a cross-walk, you are making a security trade-off: you consider the threat of getting run over versus the trouble of walking to the corner, and assess the risk of that threat happening by looking for oncoming cars. Your bodily safety is the asset you’re trying to protect. How high is the risk of getting run over and are you in such a rush that you’re willing to tolerate it, even though the threat is to your most valuable asset?
That’s a security decision. Not so hard, is it? It’s just the language that takes getting used to. Security professionals use four distinct but interrelated concepts when considering security decisions: assets, threats, risks andadversaries.
An asset is something you value and want to protect. Anything of value can be an asset, but in the context of this discussion most of the assets in question are information. Examples are you or your organization’s emails, instant messages, data files and web site, as well as the computers holding all of that information.
What You Are Protecting Against
A threat is something bad that can happen to an asset. Security professionals divide the various ways threats can hurt your data assets into six sub-areas that must be balanced against each other:
- Confidentiality is keeping assets or knowledge about assets away from unauthorized parties.
- Integrity is keeping assets undamaged and unaltered.
- Availability is the assurance that assets are available to parties authorized to use them.
- Consistency is when assets behave and work as expected, all the time.
- Control is the regulation of access to assets.
- Audit is the ability to verify that assets are secure.
Threats can be classified based on which types of security they threaten. For example, someone trying to read your email (the asset) without permission threatens its confidentiality and your control over it. If, on the other hand, an adversary wants to destroy your email or prevent you from getting it, the adversary is threatening the email’s integrity and availability. Using encryption, as described later in this guide, you can protect against several of these threats. Encryption not only protects the confidentiality of your email by scrambling it into a form that only you or your intended recipient can descramble, but also allows you to audit the emails — that is, check and see that the person claiming to be the sender is actually that person, or confirm that the email wasn’t changed between the sender and you to ensure that you’ve maintained the email’s integrity and your control over it.
The Likelihood of a Threat Actually Occuring
Risk is the likelihood that a particular threat against a particular asset will actually come to pass, and how damaged the asset would be. There is a crucial distinction between threats and risks: threats are the bad things that can happen to assets, but risk is the likelihood that specific threats will occur. For instance, there is a threat that your building will collapse, but the risk that it will really happen is far greater in San Francisco (where earthquakes are common) than in Minneapolis (where they are not).
People often over-estimate and thus over-react to the risk of unlikely threats because they are rare enough that the worst incidents are well publicized or interesting in their unusualness. Similarly, they under-estimate and under-react to more common risks. The most clichéd example is driving versus flying. Another example: when we talk to individuals about government privacy intrusions, they are often concerned about wiretapping or searches, but most people are much more at risk from less dramatic measures, like subpoenas demanding records from you or your email provider. That is why we so strongly recommend good data practices — if it’s private, don’t give it to others to hold and don’t store it, but if you do store it, protect it — while also covering more unusual circumstances, like what to do when the police show up at your door or seize your laptop.
Evaluating risk is necessarily a subjective process; not everyone has the same priorities or views threats in the same way. Many people find certain threats unacceptable no matter what the risk, because the mere presence of the threat at any likelihood is not worth the cost. In other cases, people disregard high risks because they don’t view the threat as a problem. In a military context, for example, it might be preferable for an asset to be destroyed than for it to fall into enemy hands. Conversely, in many civilian contexts, it’s more important for an asset such as email service to be available than confidential.
In his book Beyond Fear, security expert Bruce Schneier identifies five critical questions about risk that you should ask when assessing proposed security solutions:
- What assets are you trying to protect?
- What are the risks to those assets?
- How well does the security solution mitigate those risks?
- What other risks does the security solution cause?
- What costs and trade-offs does the security solution impose?
Security is the art of balancing the value of the asset you are trying to protect against the costs of providing protection against particular risks. Practical security requires you to realistically judge the actual risk of a threat in order to decide which security precautions may be worth using to protect an asset, and which precautions are absolutely necessary.
In this sense, protecting your security is a game of tradeoffs. Consider the lock on your front door. What kind of lock — or locks — should you invest in, or should you lock the door at all? The assets are invaluable — the privacy of your home and control over the things inside. The threat level is very high — you could be financially wiped out, and all of your most valuable and private information exposed, if someone broke in. The critical question then becomes: how serious is the risk of someone breaking in? If the risk is low, you probably won’t want to invest much money in a lock; if the risk is high, you’ll want to get the best locks that you can.
A critical part of assessing risk and deciding on security solutions is knowing who or what your adversary is. An adversary, in security-speak, is any person or entity that poses a threat against an asset. Different adversaries pose different threats to different assets with different risks; different adversaries will demand different solutions.
For example, if you want to protect your house from a random burglar, your lock just needs to be better than your neighbors’, or your porch better lit, so that the burglar will choose the other house. If your adversary is the government, though, money spent on a better lock than your neighbors’ would be wasted — if the government is investigating you and wants to search your house, it won’t matter how well your security compares to your neighbors. You would instead be better off spending your time and money on other security measures, like encrypting your valuable information so that if it’s seized, the government can’t read it.
Here are some examples of the kinds of adversaries that may pose a threat to your digital privacy and security:
- U.S. government agents that follow laws which limit their activities
- U.S. government agents that are willing and able to operate without legal restrictions
- Foreign governments
- Civil litigants who have filed or intend to file a lawsuit against you
- Companies that store or otherwise have access to your data
- Individual employees who work for those companies
- Hackers or organized criminals who randomly break into your computer, or the computers of companies that store your data
- Hackers or organized criminals that specifically target your computer or the computers of the companies that store your data
- Stalkers, private investigators or other private parties who want to eavesdrop on your communications or obtain access to your machines
This guide focuses on defending against threats from the first adversary — government agents that follow the law — but the information herein should also provide some help in defending against the others.
Putting it All Together
Which Threats from Which Adversaries Pose the Highest Risk to Your Assets?
Putting these concepts together, you need to evaluate which threats to your assets from which adversaries pose the most risk, and then decide how to manage the risk. Intelligently trading off risks and costs is the essence of security. How much is it worth to you to manage the risk? For example, you may recognize that government adversaries pose a threat to your webmail account, because of their ability to secretly subpoena its contents. If you consider that threat from that adversary to be a high risk, you may choose not to store your email messages with the webmail company, and instead store it on your own computer. If you consider it a low risk, you may decide to leave your email with the webmail company — trading security for the convenience of being able to access your email from any internet-connected computer. Or, if you think it’s an intermediate risk, you may leave your email with the webmail company but tolerate the inconvenience of using encryption to protect the confidentiality of your most sensitive emails. In the end, it’s up to you to decide which trade-offs you are willing to make to help secure your assets.
A Few Parting Lessons
Now that we’ve covered the critical concepts, here are a few more basic lessons in security-think that you should consider before reading the rest of this guide:
Knowledge is Power. Good security decisions can’t be made without good information. Your security tradeoffs are only as good as the information you have about the value of your assets, the severity of the threats from different adversaries to those assets, and the risk of those attacks actually happening. We’re going to try to give you the knowledge you need to identify the threats to your computer and communications security that are posed by the government, and judge the risk against possible security measures.
The Weakest Link. Think about assets as components of the system in which they are used. The security of the asset depends on the strength of all the components in the system. The old adage that “a chain is only as strong as its weakest link” applies to security, too: The system as a whole is only as strong as the weakest component. For example, the best door lock is of no use if you have cheap window latches. Encrypting your email so it won’t get intercepted in transit won’t protect the confidentiality of that email if you store an unencrypted copy on your laptop and your laptop is stolen.
Simpler is Safer and Easier. It is generally most cost-effective and most important to protect the weakest component of the system in which an asset is used. Since the weak components are much easier to identify and understand in simple systems, you should strive to reduce the number and complexity of components in your information systems. A small number of components will also serve to reduce the number of interactions between components, which is another source of complexity, cost, and risk.
More Expensive Doesn’t Mean More Secure. Don’t assume that the most expensive security solution is the best, especially if it takes away resources needed elsewhere. Low-cost measures like shredding trash before leaving it on the curb can give you lots of bang for your security buck.
There is No Perfect Security — It’s Always a Trade-Off. Set security policies that are reasonable for your organization, for the risks you face, and for the implementation steps your group can and will take. A perfect security policy on paper won’t work if it’s too difficult to follow day-to-day.
What’s Secure Today May Not Be Secure Tomorrow. It is also crucially important to continually re-evaluate the security of your assets. Just because they were secure last year or last week doesn’t mean they’re still secure!
Foreign Intelligence and Terrorism Investigations
All of the government surveillance tactics and standards discussed in previous sections relate to law enforcement investigations — that is, investigations for the purpose of gathering evidence for criminal prosecution. However, the government also engages in surveillance in order to combat foreign threats to national security. When it comes to foreign spies and terrorists, the government uses essentially the same tools — searches,wiretaps, pen/traps, subpoenas — but operates under much lower legal standards and in much greater secrecy. It’s important that you understand these foreign intelligence surveillance authorities such as the government’s access to records using National Security Letters and its wiretapping powers under the Foreign Intelligence Surveillance Act (FISA) so that you can evaluate the risk of such surveillance to you or your organization and defend against it.
National Security Letters
Imagine if the FBI could, with only a piece of paper signed by the special agent in charge of your local FBI office, demand detailed information about your private Internet communications directly from your ISP, webmail service, or other communications provider. Imagine that it could do this:
- without court review or approval
- without you being suspected of a crime
- without ever having to tell you that it happened
Further imagine that with this piece of paper, the FBI could see a wide range of private details, including:
- your basic subscriber records, including your true identity and payment information
- your Internet Protocol address and the IP address of every Web server you communicate with
- the identity of anyone using a particular IP address, username, or email address
- the email address or username of everyone you email or IM, or who emails or IMs you
- the time, size in bytes, and duration of each of your communications, and possibly even the web address of every website you visit
Finally, imagine that the FBI could use the same piece of paper to gain access your private credit and financial information — and that your ISP, bank, and any other business from which the FBI gathers your private records is barred by law from notifying you.
Now, stop imagining: the FBI already has this authority, in the form of National Security Letters. These are essentially secret subpoenas that are issued directly by the FBI without any court involvement. Thanks to the USA PATRIOT Act, the only requirement the government must meet to issue an NSL is that the FBI must certify in the letter that the information it is seeking is relevant to an authorized investigation to protect against international terrorism or clandestine intelligence activities.
The number of National Security Letters used each year is classified, but the Washington Post has reported that by late 2005, the government had on average issued 30,000 National Security Letters each year since the PATRIOT Act passed in 2001. That’s a hundredfold increase over the pre-PATRIOT numbers.
Further revelations by the FBI’s Inspector General in 2007 showed that in many cases, the FBI had failed even to meet the weak post-PATRIOT National Security Letter standards, illegally issuing so-called “exigent letters” to communications providers asking for the same information National Security Letters are used to obtain, but without meeting the minimal requirement that the requested information be relevant to an authorized terrorism or espionage investigation. EFF has since sued the Department of Justice to learn more about how the government has been abusing its National Security Letter authority.
Surveillance Under the Foreign Intelligence Surveillance Act (FISA)
The History of FISA
As stated above, the government was free to wiretap whenever it wanted to in law enforcement investigations until the Supreme Court addressed the issue in 1967, and Congress passed the Wiretap Act in 1968. Similarly, the legality of warrantless searches and wiretaps in national security investigations, as opposed to law enforcement investigations, wasn’t settled until the seventies.
In 1972, the Supreme Court ruled on the use of wiretaps in national security cases. In that case, a group of Americans protesting the Vietnam War tried to blow up their local CIA recruiting office. Investigators collected evidence against them with a wiretap but without getting a wiretap order, and argued in court that since the investigation was for national security, the president had the authority to authorize surveillance without having to go through the courts.
The Supreme Court held that the government didn’t have unlimited power to conduct surveillance without the approval of a judge just by claiming the investigation was for national security, at least when investigating domestic threats to national security (that is, threats from U.S. citizens and legal residents). It left open whether or not such warrantless surveillance was allowed when investigating foreign threats.
After this decision, and after revelations throughout the seventies that the government had been engaging in an enormous amount of unauthorized spying during the 1960s and early 1970s, Congress decided to provide a legal framework to rein in foreign intelligence investigations. The Foreign Intelligence Surveillance Act of 1978 (or “FISA”), along with later amendments to that act, created a warrant procedure for foreign intelligence investigations so that there would no longer be any foreign intelligence surveillance without court oversight.
FISA in Action
FISA requires the government to get search warrants and wiretap orders from a court even when it is investigating foreign threats to national security. However, the FISA process is different from the law enforcement processes described in earlier sections.
First, all government requests for foreign intelligence surveillance authorization are made to a secret court: the FISA court. In order to get authorization, a significant purpose of the surveillance must be to gather foreign intelligence information — information about foreign spies, foreign terrorists, and other foreign threats — instead of evidence of a crime.
Most importantly, the probable cause standard is very different. Instead of having to show probable cause that a crime is being, has been, or will be committed, the government must show that the target of the surveillance is a foreign power or an agent of a foreign power.
Also unlike law enforcement surveillance, the target is never told by the government that he/she was spied on, and every person that is served with a FISA search warrant, wiretap or pen/trap order, or subpoena is also served with a gag order forbidding them from every telling anyone about it except their lawyer.
Foreign Powers and Their Agents. So, what exactly qualifies as a foreign power or agent of a foreign power when it comes to FISA surveillance? It’s a bit unclear. The FISA law defines those terms only vaguely, and without any access to the decisions of the secret FISA court, there’s no way of telling how broadly or narrowly the definitions are being interpreted.
According to FISA, a Foreign Power is defined to include:
- Any foreign government or component of a foreign government, whether or not officially recognized by the United States
- Any “faction” of a foreign nation or nations, or any foreign-based political organization, that isn’t “substantially” composed of United States persons (“faction” and “substantially” aren’t defined; a U.S. person is a citizen or a legal resident of the U.S.)
- Any entity, like a political organization or a business, that is directed or controlled by a foreign government
- Any group engaged in, or preparing to engage in, “international terrorism.” (“International terrorism” is broadly defined as activities that (1) involve violent acts or acts dangerous to human life that are a violation of U.S. criminal laws or would be a violation if committed in the U.S., (2) appear to be intended to intimidate or coerce a civilian population, to influence the policy of a government by intimidation or coercion, or to affect the conduct of a government by assassination or kidnapping, and (3) occur totally outside the U.S., or transcend national boundaries in terms of how they are accomplished, the people they are intended to coerce or intimidate, or the place where the terrorists operate)
According to FISA, an Agent of a Foreign Power is defined to include:
- Anyone that is not a U.S. person who is an officer or employee of a foreign power
- Anyone that is not a U.S. person who engages in “clandestine intelligence activities” (spying) in the U.S. on behalf of a foreign power or any U.S. person that does the same and may be violating the law. So, if you’re not a U.S. person, you don’t have to be suspected of a crime; but even if you are a U.S. person, that suspicion doesn’t have to meet traditional probable cause standards
- Anyone, whether a U.S. person or not, who engages in or prepares for acts of international terrorism or sabotage
If you think that all sounds like very vague gobbledy-gook, you’re right. No one really knows what these terms mean other than the FISA court, which won’t release its decisions.
And it’s even worse for FISA subpoenas, which can be used to force anyone to hand over anything in complete secrecy, and which were greatly strengthened by Section 215 of the USA PATRIOT Act. The government doesn’t have to show probable cause that the target is a foreign power or agent — only that they are seeking the requested records “for” an intelligence or terrorism investigation. Once the government makes this assertion, the court must issue the subpoena.
Police at the door: FISA Orders and National Security LettersIf federal agents serve you with a FISA warrant or subpoena, or a National Security Letter, the advice given for regular warrants and subpoenas applies. However, FISA orders and National Security Letters will also come with a gag order that forbids you from discussing them. Do NOT violate the gag order. Only speak to members of your organization whose participation is necessary to comply with the order, and your lawyer. The constitutionality of FISA orders and especially National Security Letters is a matter of great dispute — in particular, several courts have found that the gag order that comes with a National Security Letter violates the First Amendment — and you may be able to successfully challenge the government’s demand in court. If you do decide to seek counsel and do not have an a lawyer of your own, you can call the lawyers at EFF.
FISA Wiretap Statistics
Like law enforcement wiretaps, FISA surveillance is relatively rare. Also like law enforcement wiretaps, however, FISA surveillance probably sweeps in the communications of a great many people. Because the information released about FISA surveillance is so limited, though, it’s impossible to gauge just how many people are affected and how many communications are intercepted. The only public data available on FISA are the numbers of applications made to, and approved by, the FISA court. And those numbers have steadily increased through the years, to the point where FISA orders now outnumber all federal and state wiretap orders combined! For example, in 2007, 2,370 applications for FISA wiretaps were granted by the FISA court, compared to 2,208 state and federal wiretaps reported in the same year. And each application can contain a request for more than one type of surveillance — for example, a wiretap, a secret search, and secret subpoenas.
Like with law enforcement wiretaps, your FISA wiretap risk is very low, as is the risk of being subjected to a secret physical search under FISA. The risk of having records about you secretly subpoenaed under FISA is much higher, but if it’s your communications records the government is after, they’re more likely to use a National Security Letter.
Privacy tip: Foreign Intelligence SurveillanceIf your organization deals with lots of non-U.S. persons or any foreign governments or foreign-based organizations, you will likely face a higher risk of foreign intelligence surveillance, and should factor that risk into your security decision-making.
The NSA Surveillance Program, the Protect America Act and the FISA Amendments Act
FISA is a dangerously weak restraint on the government’s power to secretly spy on Americans without probable cause of a crime, particularly since passage of the USA PATRIOT Act in 2001. Yet just as the Bush Administration was successfully lobbying Congress to expand its FISA surveillance authority through the USA PATRIOT Act, it was already building a new surveillance program at the National Security Agency (NSA) that would secretly ignore FISA’s limitations and spy on Americans without first going to the FISA court.
The NSA’s Surveillance Program Revealed
In a story published on December 16, 2005, the New York Times first revealed to the country that since 9/11, the NSA had regularly targeted Americans in the U.S. for electronic surveillance without first obtaining the required court orders from the FISA court. The president and his representatives quickly admitted that the Bush administration had chosen to bypass FISA as part of its “Terrorist Surveillance Program” or “TSP.” The administration claimed that the TSP was narrowly targeted at international communications — i.e., communications into and out of the country — where at least one of the parties had known links to terrorist organizations. The president made the frighteningly broad claim that because of his inherent power under the Constitution to combat foreign threats as Commander-in-Chief, he had the authority to order such warrantless surveillance regardless of FISA’s dictates or the Fourth Amendment.
However, the warrantless surveillance proved to be much broader than the “narrow and targeted” program that the president described. Further reporting by the Times and other papers made clear that the NSA’s surveillance program went far beyond the admitted “TSP.” Those news stories, along with whistleblower evidence [PDF], demonstrated that the NSA program amounted to an untargeted dragnet of millions of ordinary Americans’ domestic communications and communications records. With the cooperation of the country’s major telecommunications companies such as AT&T, the NSA had illegally gained backdoor access to critical telecommunications switching facilities and communications records databases around the nation. With that illegal access, the government was vacuuming up all of the data passing through those facilities — not only records of who communicated with whom and when but also the content of nearly every American’s private communications — as part of a vast data-mining program. In response to the mounting evidence of a dragnet surveillance program (view a summary of all of that evidence [PDF]), EFF brought suit against AT&T in 2006 — and later, in 2008, against the government itself — on behalf of ordinary AT&T customers seeking to stop the warrantless surveillance of their telephone and Internet communications. You can find out more about the progress of those lawsuits, Hepting v. AT&T and Jewel v. NSA, at our NSA Multi-District Litigation page.
The Protect America Act of 2007, the FISA Amendments Act of 2008, and the Future of the NSA’s Surveillance Program
One might expect that the revelation of a massive and illegal spying program would lead to broad bipartisan condemnation from Congress and an effort to pass legislation to provide additional protections against unbridled Executive spying. Unfortunately, that’s not what happened. Instead, the Bush administration was able to use fear of terrorism to convince Congress to pass bills authorizing surveillance programs even broader than the admitted “TSP.”
Claiming that critical intelligence about potential terrorist attacks would be lost unless FISA was immediately “modernized,” the White House succeeded in convincing Congress to pass two laws. First was the temporary Protect America Act (“PAA”) of 2007, which expired after one year. Next was the second and more-permanent FISA Amendments Act (“FAA”) of 2008. Both allowed the Executive Branch to target the communications of people outside of the U.S. for surveillance without prior FISA court approval and without demonstrating any link to terrorism. Interpreted aggressively, these statutes arguably authorized the programmatic, non-particularized dragnet surveillance of any American’s international communications, opening the door to virtually unchecked executive power to intercept your international emails and telephone calls.
In the meantime, although we don’t think that the PAA or the FAA authorizes it, there’s been no indication that the domestic dragnet, revealed by news reports and whistleblower evidence and alleged in EFF’s lawsuits, has ended. As far as we know, the NSA is still plugged into key telecommunications facilities across the country and acquiring copies of all of the communications content that flows through them, while also obtaining records detailing the communications activity of millions of ordinary Americans, in violation of FISA and the Fourth Amendment.
Considering the latest changes to the law, we strongly recommend encrypting all of your international communications traffic. As for protecting the privacy of your domestic communications, the best way to combat the NSA’s unchecked access to the nation’s communications infrastructure — short of encrypting every single communication or avoiding using telecommunications at all — is to support EFF in its litigation and lobbying efforts to stop the spying for good.
To sum up, the steps you’d take to combat FISA surveillance or national security letters are the same ones you’d take in the law enforcement context:
If you are looking for basic technical information on how to protect the privacy of your data — whether it’s on your own computer, on the wire, or in the hands of a third party — you’ve come to the right place. Although we hope you’ll have the time to review all of the information in the SSD guide, if you’re in a hurry to get to the technical details, this is where you can read articles that will explain:
Just remember: technology changes quickly. We’ll be doing our best to keep these articles updated to reflect current developments, but in the meantime, you should take the time to review information from multiple sources before making any serious security decisions.
The Internet is a global network of many individual computer networks, all speaking the same computer language, the Internet Protocol (IP). Every computer connected to the Internet has an IP address, a unique numeric identifier that can be “static”, i.e. unchanging, or may be “dynamically” assigned by your ISP, such that your computer’s address changes with each new Internet session.
More sophisticated networking protocols may be “layered” on top of the IP protocol, enabling different types of Internet communications. For instance, World Wide Web (Web) communications are transmitted via the HyperText Transfer Protocol (HTTP) and e-mails via the Simple Mail Transport Protocol (SMTP).
These additional protocols use their own types of addresses, apart from IP addresses. For example, to download a Web page, you need its Web address, known as a Uniform Resource Locator (URL) (e.g.,http://www.eff.org). To exchange e-mails, both the sender and recipient need e-mail addresses (e.g., email@example.com).
Computers that offer files for download over the Internet are called servers or hosts. For example, a computer that offers Web pages for download is called an HTTP server or Web host. Any computer may be server, client, or both, depending on the communication. The amount of data in an Internet communication is measured in bytes.
Communications to and from an Internet-connected computer occur through 65,536 different computer software “ports.” Many networking protocols have been assigned to particular port numbers by the Internet Engineering Task Force. For example, HTTP (Web) is assigned to port 80 and SMTP (e-mail) is assigned to port 25. However, any port can be used for any application, and these are only conventions.
If you want to learn more, the website How Stuff Works publishes a popular series of “Internet Basics” articles that answer questions about the nuts and bolts of the Internet.
Encryption is a technique that uses math to transform information in a way that makes it unreadable to anyone except those with special knowledge, usually referred to as a “key.” There are many applications of encryption, but some of the most important uses help protect the security and privacy of files on your computer, information passing over the Internet, or left sitting in a file on someone else’s computer. If encryption is used properly, the information should only be readable by you and people that receive the key from you. Encryption provides a very strong technical protection against many kinds of threats — and this protection is often easy to obtain.
How Does Encryption Work?
What do you need to know about how encryption works? Surprisingly little. Encryption is conceptually similar to the “secret codes” that children learn about and use to communicate. If you’ve ever spoken in pig Latin or used a decoder ring, you’ve used very simple encryption techniques on a message. Again, the idea is to take a normal human-readable message (often called the plaintext message) and transform it into an incomprehensible format that can only become comprehensible again to someone with secret knowledge:
Plaintext message + Encryption algorithm + Key = Scrambled message
Decryption algorithm + Key + Scrambled message = Plaintext Message
Your Little Brother’s Cryptography. A simple encryption system would be to change each letter in your message to a set number of letters later in the alphabet. The specific number of spaces you move down the alphabet for each letter is the secret key. If the key is two, A becomes C, B becomes D, C becomes E, etc. Using that encryption system, the plaintext message “INSECURE” would become “KPUGEWTG.”
How is Encryption Applied?
Although the mechanics of encryption can be explained by the “decoder ring” analogy, the modern practice of using encryption has been accurately described as using a very resilient envelope for your messages. Most unencrypted data transmitted online is accessible to the servers passing off the information. Conversely, using encryption puts your online communications in a “steel envelope” — they can’t be read in the course of delivering the message to the recipient and are extremely resistant to tampering.
Modern encryption is very difficult to break, using very complex mathematics to scramble information and ensure that only people possessing the right key can unscramble it. In many cases you can get major security benefits from encryption without a detailed understanding of how it works. Some software implements very convenient, fully automated encryption features which may simply require that you turn them on.
For instance, when a website is configured properly, web browsers can use SSL encryption to protect the privacy of information you send to or receive from a web server. This is most often used to protect log-in passwords and financial data. Using a browser’s SSL encryption can be as simple as accessing a site with the https scheme instead of the http scheme (for instance, https://www.eff.org/ instead of http://www.eff.org/); the browser typically takes care of all the details behind the scenes.
Why Is Encryption Important?
Encryption plays an important role in mitigating risk related to the many threats listed in this guide. If sensitive information stored on your computer is encrypted, it will take a secret key to decode it. If sensitive information en route to others is encrypted, only someone that knows the secret key can read what it says. When you encrypt sensitive information and it ends up logged by others in the course of communicating online, encryption keeps those without the secret key from knowing the contents of the message.
Most of the Defensive Technology articles in this guide will cover practical ways to apply encryption to particular communications (like email) or particular applications (like web browsers).
Encryption is absolutely essential to maintaining information security. Moreover, modern computers are powerful enough that we can aim to make encryption of our communications and data routine, not just reserving encryption for special occasions or particularly sensitive information.
For More About Encryption
Many encryption tools can be used successfully without much beyond a conceptual understanding. We explain how to use many of these well-developed tools in other parts of this guide.
However, be aware that while encryption is a powerful tool and is critical to information security, it has limitations — particularly if it is not being used correctly. Learning more about encryption and its limitations can help ensure that you’re using it properly and getting protection against as many kinds of attacks as possible.
Web browsers are software on your machine that communicate with servers or hosts on the Internet. Using a web browser causes data to be stored on your computer and logs to be stored on the web servers you visit, and frequently transmits unencrypted information.
Until you have understood the mechanisms by which this occurs — and taken steps to prevent them — it is best to assume that anything you do with a web browser could be recorded by your own machine, by the web servers you’re communicating with, or by any adversary that is able to monitor your network connection.
Controlling and Limiting the Logs Kept by Your Browser
Web browsers often retain a large amount of information about the way they are used. A browser typically keeps a history of the web pages it visits. Browsers also often retain cached copies of the pages you’ve visited, information about which accounts you log into on web servers, names and other data you enter into web forms, and cookies that record preferences and link your browser to records on third party web servers. Fortunately, browsers also include features for managing these records. In general, the features are getting better, so it’s getting easier to control browser records.
For example, here are the stored data privacy settings pages for Firefox, the free web browser:
For each type of information your browser stores, you can either set it to not collect it at all, set it to delete within a certain span of days, set it to delete when you quit the browser, or press “clear” to manually erase the data. Or you can “clear all” of the info — all the data your browser’s been keeping on you.
Apple’s Safari browser also has an easy one-click option to clear everything. Just select “Reset Safari” from the “Safari” pull-down menu and you’ll get this option:
Controlling and Limiting the Logs Kept By Web Servers
Web servers usually see and retain a large amount of information about what you do when you surf to them. For instance, if you type any information into a form on a web page (such as a search engine), the server will record not only what you sent it, but also information that might identify you: your IP address, the browser and operating system you are using, whether you followed a link from another web page to get to the page, what that previous site/page was, your account if you are logged in to the site, and cookies that were created when you previously looked at pages on the site.
Web Privacy is Hard
If you use a particular website a lot, the chances are that it is going to end up retaining a huge amount of information about you. To get a sense of the kinds of information, and what needs to be done to prevent them from being aggregated, read our white paper on search privacy. Although that document primarily discusses search engines, the issues to consider for other kinds of sites are similar.
Cookies are pieces of information that a web site can send to your browser. If your browser “accepts” them, they will be sent back to the site every time the browser accepts a page, image or script from the site. A cookie set by the page/site you’re visiting is a “second party” cookie. A cookie set by another site that’s just providing an image or script (an advertiser, for instance), is called a “third party” cookie.
Cookies are the most common mechanisms used to record the fact that a particular visitor has logged in to an account on a site, and to track the state of a multi-step transaction such as a reservation or shopping cart purchase. As a result, it is not possible to block all cookies without losing the ability to log into many sites and perform transactions with others.
Unfortunately, cookies are also used for other purposes that are less clearly in users’ interests, such as recording their usage of a site over a long period of time, or even tracking and correlating their visits to many separate sites (via cookies associated with advertisements, for instance).
With recent browsers, the cookie setting that offers users the most pragmatic trade-off between cookie-dependent functionality and privacy is to only allow cookies to persist until the user quits the browser (also known as only allowing “session cookies”).
You can enable this in the “Privacy” tab of Firefox 3’s “Preferences” pane:
Unfortunately, if you only quit your browser entirely once every week or two, web sites will still collect a huge amount of information about your habits, such as the IP addresses you use at home, at work, at friends’ houses and at Internet cafes. However, the “Incognito” mode offered by Google’s Chrome browser and the “InPrivate” mode offered by Internet Explorer 8 are signs that in future browsers may offer more convenient ways to limit cookie tracking.
Sophisticated users can configure their browser to manually decide whether each site they visit is allowed to set cookies. This may have good privacy outcomes, such as allowing session cookies for sites the user logs in to or purchases things from, but not any other sites. But it requires a lot of work. A certain amount of debugging may also be required for situations where sites are poorly designed and fail to function without certain third-party cookies.
Recent Cookie-Like “Features” in Web Browsers
In addition to the regular cookies that web browsers send and receive, and which users have begun to be aware of and manage for privacy, companies have continued to implement new “features” which behave like cookies but which aren’t managed in the same way. Adobe has created “Local Stored Objects” (also known as “Flash Cookies”) as a part of its Flash plug-ins; Mozilla has incorporated a feature called “DOM storage” in recent versions of Firefox. Web sites could use either or both of these in addition to cookies to track visitors. We recommend that users take steps to prevent this.
Managing Mozilla/Firefox DOM Storage Privacy. If you use a Mozilla browser, you can disable DOM Storage pseudo-cookies by typing about:config into the URL bar. That will bring up an extensive list of internal browser configuration options. Type “storage” into the filter box, and press return. You should see an option called dom.storage.enabled. Change it to “false”.
Managing Adobe Flash Privacy. Adobe lists advice on how to disable Flash cookies here. There are some problems with the options Adobe offers (for instance, there is no “session only” option), so it’s probably best to globally set Local Stored Object space to 0 and only change that for sites which you’re willing to have tracking you. On the Linux version of Adobe’s Flash plugin there doesn’t seem to be a way set the limit to 0 for all sites — consider donating or contributing to the Gnash project to give users an alternative to Adobe’s privacy-unfriendly design decisions.
Aside from being an annoying medium for advertising, Flash poses other kinds of privacy and security risks. Some people choose not to use Flash at all (using other tools like youtube-dl for watching Youtube videos). Others install a Flash management browser plugin like FlashBlocker. Unfortunately, while FlashBlocker makes surfing the web a more peaceful experience, it does not protect you from being tracked by Flash cookies or from exposure to other flash-based security risks.
Whenever your browser fetches a page, image or script from a website, you should expect the website to record the IP address of the computer you’re using. Your ISP, or anybody with the power to subpoena your ISP, could tie those records to the Internet account subscription you are connected through. Use Tor (or a proxy server, which is faster but less secure) if you wish to prevent these records from being created.
Privacy on the wire
Most sites on the web are accessed using the unencrypted HTTP protocol. HTTP is susceptible to eavesdropping, and even to intermediaries that might set out to modify the pages a browser is fetching.
HTTPS is a more secure alternative to HTTP. HTTPS encrypts pages, and attempts to ensure three things: (1) that third parties cannot see the contents of the page; (2) that the page cannot be modified by third parties; (3) that the page was really sent by the web server listed in the URL bar.
Unfortunately, a web server must be configured to support HTTPS properly before you can use it. If there is a site you were planning to send sensitive information to, ensure that you are using HTTPS. If a site doesn’t support HTTPS, don’t send sensitive information to it.
Some Notes on Using HTTPS
Check three indicators to ensure that you’re at an HTTPS page: (1) the URL begins with https://; (2) there is a lock icon in the corner of the browser; and (3) the URL/location bar is colored.
If you receive a warning about certificates, or a see broken lock icon, you should assume that any of the security properties of the page could be broken. Contact the site’s webmaster and have them fix the problem before sending any sensitive information to the site.
The act of using email stores data on your machines, transmits data over the network, and stores data on third party machines.
Locally Stored Data
The usual measures apply to managing the copies of emails (both sent and received) that are kept on your own machines. Encrypt your drives and decide upon and follow an appropriate data deletion policy.
Data on the Wire
Email usually travels through a number of separate hops between the sender and receiver. This diagram illustrates the typical steps messages might travel through, the transmission protocols used for those steps, and the available types of encryption for those steps.
End-to-End Encryption of Specific Emails
Encrypting emails all the way from the sender to the receiver has historically been difficult, although the tools for achieving this kind of end-to-end encryption are getting better and easier to use. Pretty Good Privacy (PGP) and its free cousin GNU Privacy Guard (GnuPG) are the standard tools for doing this. Both of these programs can provide protection for your email in transit and also protect your stored data. Major email clients such as Microsoft Outlook and Mozilla Thunderbird can be configured to work smoothly with encryption software, making it a simple matter of clicking a button to sign, verify, encrypt and decrypt email messages.
The great thing about end-to-end encryption is that it ensures that the contents of your emails will be protected not only against interception on the wire, but also against some of the threats to the contents of copies of your emails stored on your machine or third party machines.
There are two catches with GnuPG/PGP. The first is that they only work if the other parties you are corresponding with also use them. Inevitably, many of the people you exchange email with will not use GPG/PGP, though it can be deployed amongst your friends or within an organization.
The second catch is that you need to find and verify public keys for the people you are sending email to, to ensure that eavesdroppers cannot trick you into using the wrong key. This trickery is known as a “man in the middle” attack.
Probably the easiest way to start using GnuPG is to use Mozilla Thunderbird with the Enigmail plugin. You can find the quick start guide for installing and configuring Enigmail here.
Server-to-Server Encrypted Transit
After you press “send”, emails are typically relayed along a chain of SMTP mail servers before reaching their destination. You can use your mail client to look at the headers of any email you’ve received to see the chain of servers the message traveled along. In most cases, messages are passed between mail servers without encryption. But there is a standard called SMTP over TLS which allows encryption when the sending and receiving servers for a given hop of the chain support it.
If you or your organization operates a mail server, you should ensure that it supports TLS encryption when talking to other mail servers. Consult the documentation for your SMTP server software to find out how to enable TLS.
Client-to-Mail Server Encryption
If you use POP or IMAP to fetch your email, make sure it is encrypted POP or IMAP. If your mail server doesn’t support the encrypted version of that protocol, get your service provider or systems administrator to fix that.
If you use a webmail service, ensure that you only access it using HTTPS rather than HTTP. Hushmail.com is a webmail service provider that always uses HTTPS, and also offers some end-to-end encryption facilities (though they are not immune to warrants).
Many webmail service providers only use HTTPS for the login page, and then revert to HTTP. This isn’t secure. Look for an account configuration option (or a browser plugin) to ensure that your webmail account always uses HTTPS. In Gmail, for instance, you can find this option in the “general” tab of the settings page:
If you can’t find a way to ensure that you only see your webmail through https, switch to a different web mail provider.
Data Stored on Second- and Third-Party Machines
There are two main reasons why your emails will be stored on computers controlled by third parties.
Storage by your Service Provider
If you don’t run your own mail server, then there is a third party who obtains (and may store) copies of all of your emails. This would commonly be an ISP, an employer, or a webmail provider. Copies of messages will also be scattered across computers controlled by the ISPs, employers and webmail hosts of those you correspond with.
Make sure your email software is configured so that it deletes messages off of your ISP’s mail server after it downloads them. This is the most common arrangement if you’re using POP to fetch your email, but it is common for people to use IMAP or webmail to leave copies of messages on the server.
If you use webmail or IMAP, make sure you delete messages immediately after you read them. Keep in mind that with major webmail services, it may be a long time – maybe a matter of months – before the message is really deleted, regardless of whether you still have access to it or not. With smaller IMAP or webmail servers, it is possible that forensically accessible copies of messages could be subpoenaed years after the user deleted them.
The content of PGP/GnuPG encrypted emails will not be accessible through these third parties, although the email headers (such as the To: and Subject: lines) will be.
Running your own mail server with an encrypted drive, or using end-to-end encryption for sensitive communications, are the best ways of mitigating these risks.
Storage by Those You Correspond With
Most people and organizations save all of the email they send and receive. Therefore, almost every email you send and receive will be stored in at least one other place, regardless of the practices and procedures you follow. In addition to the personal machine of the person you sent/received the message to/from, copies might be made on their ISP or firm’s mail or backup servers. You should take these copies into consideration, and if the threat model you have for sensitive communications includes an adversary that might gain access to those copies, then you should either use PGP to encrypt those messages, or send them by some means other than email. Be aware that even if you use PGP, those you communicate with could be subject to subpoenas or requests from law enforcement to decrypt your correspondence.
End-to-End Email Encryption
Email encryption is a topic that could fill a book, and has: see Bruce Schneier’s book Email Security: How to Keep Your Electronic Messages Private. While this book is somewhat out of date (it refers to old versions of software), the concepts it introduces are essential.
Instant Messaging (IM)
Instant messaging is a convenient way to communicate with people online. In privacy terms, it’s a bit better and easier to secure than email but in some situations a telephone call will offer you better privacy.
Instant messaging software creates data stored on your computer (logs of your communications), transmits communications over the network (the messages traveling back and forth), and leaves communications stored on other computers (logs kept by the people you talk to, and sometimes logs kept by the IM provider).
If you use IM without taking special precautions, you can assume that all of these records will be available to adversaries. The easiest way for an adversary to obtain the contents of your communications is from you, your correspondent, or your service provider, if any of those parties logs (stores) the messages. The more difficult way is to intercept the messages as they travel over the network.
Encrypt Your Instant Messaging Conversations as They Travel
To protect messages from interception as they travel over the network, you need to use encryption. Fortunately, there is an excellent instant messaging encryption system called OTR (Off The Record). Confusingly, Google has a different instant messaging privacy feature which is also called “Off The Record”. To disambiguate them, this page will talk bout “OTR encryption” and “Google OTR”. It’s actually possible to be using them both at the same time.
If you and the person you are talking to both use OTR encryption, you have excellent protection for communications on the network, and you will prevent your IM provider from storing the content of your communications (though they may still keep records of who you talk to).
The easiest way to use OTR encryption is to use Pidgin or Adium for your IMs (Pidgin is a program that will talk to your friends over the MSN, Yahoo!, Google, Jabber, and AIM networks; Adium is a similar program specifically for Mac OS X). If you’re using Pidgin, install the the OTR encryption plugin for that client. Adium comes with OTR built in.
With OTR encryption installed, you still need to do a few things for network privacy:
- Read and understand OTR encryptions’s information.
- Make sure the people you are talking to also use OTR encryption, and make sure it’s active. (In Pidgin, check for OTR:private or OTR:unverfied in the bottom right corner.)
- Follow OTR encryption’s instructions to “Confirm” any person you need to have sensitive conversations with. This reduces the risk of an interloper (including the government with a warrant) being able to trick you into talking to them instead of the person you meant to talk to. Recent versions of OTR encryption allow you to do this just by agreeing on a shared secret word that you both have to type (“what was the name of the friend who introduced us?”). Older versions required that both users check that their client reported the right fingerprint for the other client.
Configure Your IM Client to use SSL/TLS
This step is complementary to using OTR encryption. It will prevent someone watching the network from seeing who you are chatting to, and will offer partial protection of your chats even if the other party isn’t using OTR.
If you are using Pidgin, you can ensure SSL is enabled by going to Manage Accounts, selecting Modify for an account, selecting the Advanced tab, and ticking Require SSL/TLS.
Understand and Control IM Logging on Your Machine
To protect the privacy of your IM conversations, you will need to decide what to do about logs kept on your computer. You have three choices:
- Configure your IM client to not keep logs
- Encrypt your hard disk
- Accept the risk that anyone who has access to your computer can read your old messages
If at some point you decide to configure your IM client not to keep logs, you may want to go back and delete previous logs using Secure Deletion software.
Be Aware of Logging on Others’ Machines
As noted above, using OTR encryption will ensure that your IM service provider should be unable to log the contents of your communications. They will, however, be in a position to record who you talk to, and possibly record the timing and length of the messages you exchange.
OTR encryption does not stop the people you are talking to from logging your conversations. Unless you trust that they have disabled logging in their client or that they encrypt their hard disk and will not turn over its contents, you should assume that an adversary could obtain records of your conversations from the other party, either voluntarily or through subpoena or search.
Google OTR is a feature of the Google instant messaging service that allows you to request that neither Google nor the people your talk to should be able to log your conversations. Unfortunately, there is no plausible enforcement mechanism for this feature. The people you talk to could be using a different IM client (like Pidgin or Adium) that can log regardless of whether Google OTR is enabled — or they could take screenshots of your conversations. Your client might be able to tell you whether they are using a client that follows the OTR rules (such as Gmail or Gchat), but that won’t tell you whether they are taking screenshots. The bottom line is that Google OTR is nice in theory but insecure in practice. Turn it on, but don’t expect it to work if the other party uses a non-Google client or actively wants to record the converstion.
Wireless networking is now a ubiquitous means of connecting computers to each other and to the Internet. The primary privacy concern with Wi-Fi is the interception of the communications you send over the air. In some cases, wireless routers might also store a small amount of information about your computer, such as its name and the unique number assigned to its networking card (MAC address).
Wireless networks are particularly vulnerable to eavesdropping — in the end, “wireless” just means “broadcasting your messages over the radio,” and anyone can intercept your wireless signal unless you use encryption. Listening in on unencrypted Wi-Fi communications is easy: almost any computer can do it with simple packet-sniffing software. Special expertise or equipment isn’t necessary.
Even worse, the legal protections for unencrypted wireless communications are unclear. Law enforcement may be able to argue that it does not need a wiretap order to intercept unencrypted wi-fi communications because there is an exception to the rules requiring such orders when the messages that are being intercepted are “readily accessible to the public.” Basically, any communication over the radio spectrum that isn’t transmitted by your phone company and isn’t scrambled or encrypted poses a privacy risk.
Encrypting a Wireless Network
If you want to protect your wireless communications from the government or anyone else, you must use encryption! Almost all wireless Internet access points come with WEP (Wired Equivalent Privacy) or WPA (Wi-Fi Protected Access) encryption software installed to encrypt the messages between your computer and the access point, but you have to read the manual and figure out how to use it. WEP is not great encryption (and we recommend strong, end-to-end encryption for sensitive communications regardless of the transmission medium), and practiced hackers can defeat it very quickly, but it’s worth the trouble to ensure that your communications will be entitled to the legal protections of the Wiretap Act. WPA is much stronger than WEP, but it still only covers the first step your packets will take across the Internet.
When Using Open Wi-Fi
If you’re using someone else’s “open” — unencrypted — wireless access point, like the one at the coffee shop, you will have to take care of your own encryption using the tools and methods described in other sections. Toris especially useful for protecting your wireless transmissions. If you don’t use Tor, and even if you do, you should also always use application-level encryption over open wireless, so no one can sniff your passwords.
Because of the threat of password sniffing, it is crucially important that you do not use the same password for all your accounts! For example, http://www.nytimes.com/ requires a username and password to log in, but the site does not use encryption. However, web sites for banks, like https://www.wellsfargo.com/, always use encryption due to the sensitive nature of the transactions people make with banks. If you use the same passwords for the two sites, an eavesdropper could see your unencrypted password traveling to the newspaper site, and guess that you were using the same password for your bank account.
Tor is another encryption tool that can help you protect the confidentiality of your communications. Tor is a free, relatively easy to use tool primarily designed to protect your anonymity on-line. But it also has the side benefit of encrypting your communications for some of their journey across the Internet.
How Tor Works
Using Tor can help you anonymize web browsing and publishing, instant messaging, IRC, SSH, and many other applications. The information you transmit is safer when you use Tor, because communications are bounced around a distributed network of servers, called onion routers. This provides anonymity, since the computer you’re communicating with will never see your IP address — only the IP address of the last Tor router that your communications travelled through.
Tor helps to defend against traffic analysis by encrypting your communications multiple times and then routing them through a randomly selected set of intermediaries. Thus, unless an eavesdropper can observe all traffic to and from both parties, it will be very hard to determine your IP address. The idea is similar to using a twisty, hard-to-follow route in order to throw off somebody who is tailing you, and then periodically erasing your footprints.
To create a private network pathway with Tor, Alice’s Tor client first queries a global directory to discover where on the Internet all the Tor servers are. Then it incrementally builds a circuit of encrypted connections through servers on the network. The circuit is extended one hop at a time, and each server along the way knows only which server gave it data and which server it is giving data to. No individual server ever knows the complete path that a data packet has taken. The Tor software on your machine negotiates a separate set of encryption keys for each hop along the circuit to ensure that each hop can’t trace these connections as they pass through.
Due to the way Alice’s Tor client encrypted her data, each node in the circuit can only know the IP addresses of the nodes immediately adjacent to it. For example, the first Tor server in the circuit knows that Alice’s Tor client sent it some data, and that it should pass that data on to the second Tor server. Similarly, Bob knows only that it received data from the last Tor server in the circuit — Bob has no knowledge of the true Alice.
For efficiency, the Tor software uses the same circuit for connections that happen within the same ten-minute period. Later requests are given a new circuit, to keep people from linking your earlier actions to the new ones.
Tor’s primary purpose is to frustrate traffic analysis, but as a by-product of how it works, Tor’s encryption provides strong protection for the confidentiality of the content of messages as well. If an eavesdropper wiretaps Alice’s network link, all she’ll see is encrypted traffic between Alice and her first Tor server — a great feature. If the eavesdropper wiretaps Bob’s network link, she can see the unencrypted content Alice sent to Bob — but it may be very hard indeed for her to link the content to Alice!
You can learn about Tor, find easy installation instructions, and download the software at http://www.torproject.org. There you will also find instructions on how to easily “Torify” all kinds of different applications, including web browsers and instant messaging clients.
What Tor Won’t Defend You Against
Tor won’t defend you against Malware. If your adversary can run programs on your computer, it’s likely that they can see where you are and what you’re doing with Tor.
If you’ve installed Tor on your computer but are using applications that don’t understand how to use it, or aren’t set up to use it, you won’t receive protection while using those applications.
Tor may not defend you against extremely resourceful and determined opponents. Tor is believed to work quite well at defeating surveillance from one or a handful of locations, such as surveillance by someone on your wireless network or surveillance by your ISP. But it may not work if someone can surveil a great many places around the Internet and look for patterns across them.
If you aren’t using encryption with the actual servers you’re communicating with (for instance, if you’re using HTTP rather than HTTPS), the operator of an “exit node” (the last Tor node in your path) could read all your communications, just the way your own ISP can if you don’t use Tor. Since Tor chooses your path through the Tor network randomly, targeted attacks may still be difficult, but researchers have demonstrated that a malicious Tor exit node operator can capture a large amount of sensitive unencrypted traffic. Tor node operators are volunteers and there is no technical guarantee that individual exit node operators won’t spy on users; anyone can set up a Tor exit node.
These and related issues are discussed in more detail at here.
Malware is a catch-all term referring to software that runs on a computer and operates against the interests of the computer’s owner. Computer viruses, worms, trojan horses, “spyware”, rootkits and key loggers are often cited as subcategories of malware. Note that some programs may belong to more than one of those categories.
How Does Malware Get Onto a Computer?
Some malware is spread by exploiting vulnerabilities in operating systems or application software. These vulnerabilities are design or programming errors in software that can allow a clever programmer to trick the defective software into giving someone else control. Unfortunately, such vulnerabilities have been found in a wide variety of mainstream software, and more are detected all the time — both by those trying to fix the vulnerabilities and by those trying to exploit them.
Another common vector by which malware spreads is to trick the computer user into running a software program that does something the user wouldn’t have wanted. Tricking the user is a pretty powerful way to take over a computer, because the attacker doesn’t have to depend on finding a serious weakness in mainstream software. It is especially difficult to be sure that computers shared by several users, or a computer in a public place such as a library or Internet café, are not compromised. If a single user is tricked into running a malware installer, every subsequent user, no matter how cautious, could be at risk. Malware written by sophisticated programmers generally leaves no immediately visible signs of its presence.
What is Malware Capable of?
Malware is extremely bad news from a security and privacy perspective. Malware may be capable of stealing account details and passwords, reading the documents on a computer (including encrypted documents, if the user has typed in the password), defeating attempts to access the Internet anonymously, taking screenshots of your desktop, and hiding itself from other programs. Malware is even capable of using your computer’s microphone, webcam, or other peripherals against you.
The chief limitation in malware’s capability is that the author needs to (1) have anticipated the need for the malware to do something, (2) spent a substantial amount of effort programming the malicious feature, testing that it works and is robust on numerous different versions of an operating system, and (3) be free of legal or other restrictions preventing the implementation of the feature.
Unfortunately, a black market has appeared in recent years that sells malware customized for various purposes. This has reduced the obstacles listed in category (2) above.
The most alarming feature of malware is that, once installed, it can potentially nullify the benefits of other security precautions. For example, malware can be used to bypass the protections of encryption software even if this software is otherwise used properly. On the other hand, the majority of malware is mainly designed to do other things, like popping up advertisements or hijacking a computer to send spam.
Is Malware Infection Likely?
Nobody knows how many computers are infected with malware, but informed estimates range from 40% to almost 90% of computers running Windows operating systems. Infection rates are lower for MacOS and Linux systems, but this is not necessarily because Windows is an easier target. Indeed, recent versions of Windows are much improved in security. Rather, more malware authors target Windows machines because an effective attack will give them control of more computers.
The risk that any given computer is infected with malware is therefore quite high unless skilled computer security specialists are putting a substantial amount of effort into securing the system. With time, any machine on which security updates are not installed promptly is virtually guaranteed to become infected. It is however overwhelmingly likely that the malware in question will be working on obtaining credit card numbers, obtaining eBay account passwords, obtaining online banking passwords, sending spam, or launching denial of service attacks, rather than spying on specific individuals or organizations.
Infection by malware run by U.S. law enforcement or other governmental agencies is also possible, though vastly less likely. There have been a handful of cases in which it is known that warrants were obtained to install malware to identify a suspect or record their communications (see the section on CIPAV below). It is unlikely that U.S. government agencies would use malware except as part of significant and expensive investigations.
How Can You Reduce the Risk of Malware Infection?
Currently, running a minority operating system significantly diminishes the risk of infection because fewer malware applications have been targeted at these platforms. (The overwhelming majority of existing malware targets only a single particular operating system.)
Vulnerabilities due to software defects are difficult to mitigate. Installing software updates promptly and regularly can ensure that at least known defects are repaired.
Not installing (or running) any software of unknown provenance is an important precaution to avoid being tricked into installing malware. This includes, for example, software applications advertised by banner ads or pop-ups, or distributed by e-mail (even if disguised as something other than a computer program). Recent operating systems attempt to warn users about running software from an unknown source; these security warnings serve an important purpose and should not be casually ignored. Strictly limiting the number of users of a computer containing sensitive information can also be helpful. Notably, some malware targets children, including malicious code along with downloadable video games. (Of course, computer users of any age can be tricked into installing malware!)
On Windows, regularly running antivirus and antispyware software can remove a large proportion of common malware. However, this software is not effective against all malware, and must be regularly updated. Since anti-malware software is created by researching malware discovered “in the wild,” it’s also probably ineffective against uncommon, specially-targeted malware applications that aim to infect only a few specific computers rather than a large population on the Internet.
CIPAV: An Example of Malware Use for Law Enforcement
A CIPAV is an FBI acronym which stands for Computer and Internet Protocol Address Verifier. CIPAVs are a type of malware intended to identify people who are hiding their identity using proxy servers, bot nets, compromised computers or anonymity networks like Tor. A small amount is known about them as a result of published documents from cases in which they were used. CIPAVs may include use of browser exploits to run software on a computer regardless of how many steps of indirection are present between the attacking server and the user.
Malware Risk Assessment
Ubiquitous malware poses a threat to all computer users. The seriousness of the threat varies greatly. For some users, it is sufficient to install operating system updates regularly and utilize caution in running software found on the web. For organizations that face a high risk of being specifically targetted by a malware author, it is advisable to find computer security experts to defend their computers — or better yet, to simply avoid using networked computers for their most sensitive activities.
This article discusses privacy implications of cell phones and other devices that communicate with large scale wireless voice and data networks