Operating Systems Design and Implementation (3rd Edition)

5.4. Security

File systems generally contain information that is highly valuable to their users. Protecting this information against unauthorized usage is therefore a major concern of all file systems. In the following sections we will look at a variety of issues concerned with security and protection. These issues apply equally well to timesharing systems as to networks of personal computers connected to shared servers via local area networks.

5.4.1. The Security Environment

People frequently use the terms "security" and "protection" interchangeably. Nevertheless, it is frequently useful to make a distinction between the general problems involved in making sure that files are not read or modified by unauthorized persons, which include technical, administrative, legal, and political issues on the one hand, and the specific operating system mechanisms used to provide security, on the other. To avoid confusion, we will use the term security to refer to the overall problem, and the term protection mechanisms to refer to the specific operating system mechanisms used to safeguard information in the computer. The boundary between them is not well defined, however. First we will look at security to see what the nature of the problem is. Later on in the chapter we will look at the protection mechanisms and models available to help achieve security.

Security has many facets. Three of the more important ones are the nature of the threats, the nature of intruders, and accidental data loss. We will now look at these in turn.

Threats

From a security perspective, computer systems have three general goals, with corresponding threats to them, as listed in Fig. 5-22. The first one, data confidentiality, is concerned with having secret data remain secret. More specifically, if the owner of some data has decided that these data are only to be made available to certain people and no others, the system should guarantee that release of the data to unauthorized people does not occur. As a bare minimum, the owner should be able to specify who can see what, and the system should enforce these specifications.

The second goal, data integrity, means that unauthorized users should not be able to modify any data without the owner's permission. Data modification in this context includes not only changing the data, but also removing data and adding false data as well. If a system cannot guarantee that data deposited in it remain unchanged until the owner decides to change them, it is not worth much as an information system. Integrity is usually more important than confidentiality.

The third goal, system availability, means that nobody can disturb the system to make it unusable. Such denial of service attacks are increasingly common. For example, if a computer is an Internet server, sending a flood of requests to it may cripple it by eating up all of its CPU time just examining and discarding incoming requests. If it takes, say, 100µsec to process an incoming request to read a Web page, then anyone who manages to send 10,000 requests/sec can wipe it out. Reasonable models and technology for dealing with attacks on confidentiality and integrity are available; foiling denial-of-services attacks is much harder.

Another aspect of the security problem is privacy: protecting individuals from misuse of information about them. This quickly gets into many legal and moral issues. Should the government compile dossiers on everyone in order to catch X-cheaters, where X is "welfare" or "tax," depending on your politics? Should the police be able to look up anything on anyone in order to stop organized crime? Do employers and insurance companies have rights? What happens when these rights conflict with individual rights? All of these issues are extremely important but are beyond the scope of this book.

Intruders

Most people are pretty nice and obey the law, so why worry about security? Because there are unfortunately a few people around who are not so nice and want to cause trouble (possibly for their own commercial gain). In the security literature, people who are nosing around places where they have no business being are called intruders or sometimes adversaries. Intruders act in two different ways. Passive intruders just want to read files they are not authorized to read. Active intruders are more malicious; they want to make unauthorized changes. When designing a system to be secure against intruders, it is important to keep in mind the kind of intruder one is trying to protect against. Some common categories are

1. Casual prying by nontechnical users. Many people have personal computers on their desks that are connected to a shared file server, and human nature being what it is, some of them will read other people's electronic mail and other files if no barriers are placed in the way. Most UNIX systems, for example, have the default that all newly created files are publicly readable.

---

[Page 528]

2. Snooping by insiders. Students, system programmers, operators, and other technical personnel often consider it to be a personal challenge to break the security of the local computer system. They often are highly skilled and are willing to devote a substantial amount of time to the effort.

3. Determined attempts to make money. Some bank programmers have attempted to steal from the bank they were working for. Schemes have varied from changing the software to truncate rather than round interest, keeping the fraction of a cent for themselves, to siphoning off accounts not used in years, to blackmail ("Pay me or I will destroy all the bank's records.").

4. Commercial or military espionage. Espionage refers to a serious and well-funded attempt by a competitor or a foreign country to steal programs, trade secrets, patentable ideas, technology, circuit designs, business plans, and so forth. Often this attempt will involve wiretapping or even erecting antennas directed at the computer to pick up its electromagnetic radiation.

It should be clear that trying to keep a hostile foreign government from stealing military secrets is quite a different matter from trying to keep students from inserting a funny message-of-the-day into the system. The amount of effort needed for security and protection clearly depends on who the enemy is thought to be.

Malicious Programs

Another category of security pest is malicious programs, sometimes called malware. In a sense, a writer of malware is also an intruder, often with high technical skills. The difference between a conventional intruder and malware is that the former refers to a person who is personally trying to break into a system to cause damage whereas the latter is a program written by such a person and then released into the world. Some malware seems to have been written just to cause damage, but some is targeted more specifically. It is becoming a huge problem and a great deal has been written about it (Aycock and Barker, 2005; Cerf, 2005; Ledin, 2005; McHugh and Deek, 2005; Treese, 2004; and Weiss, 2005)

The most well known kind of malware is the virus. Basically a virus is a piece of code that can reproduce itself by attaching a copy of itself to another program, analogous to how biological viruses reproduce. The virus can do other things in addition to reproducing itself. For example, it can type a message, display an image on the screen, play music, or something else harmless. Unfortunately, it can also modify, destroy, or steal files (by e-mailing them somewhere).

Another thing a virus can do is to render the computer unusable as long as the virus is running. This is called a DOS (Denial Of Service) attack. The usual approach is to consume resources wildly, such as the CPU, or filling up the disk with junk. Viruses (and the other forms of malware to be described) can also be used to cause a DDOS (Distributed Denial Of Service) attack. In this case the virus does not do anything immediately upon infecting a computer. At a predetermined date and time thousands of copies of the virus on computers all over the world start requesting web pages or other network services from their target, for instance the Web site of a political party or a corporation. This can overload the targeted server and the networks that service it.

Malware is frequently created for profit. Much (if not most) unwanted junk e-mail ("spam") is relayed to its final destinations by networks of computers that have been infected by viruses or other forms of malware. A computer infected by such a rogue program becomes a slave, and reports its status to its master, somewhere on the Internet. The master then sends spam to be relayed to all the e-mail addresses that can be gleaned from e-mail address books and other files on the slave. Another kind of malware for profit scheme installs a key logger on an infected computer. A key logger records everything typed at the keyboard. It is not too difficult to filter this data and extract information such as username password combinations or credit card numbers and expiration dates. This information is then sent back to a master where it can be used or sold for criminal use.

Related to the virus is the worm. Whereas a virus is spread by attaching itself to another program, and is executed when its host program is executed, a worm is a free-standing program. Worms spread by using networks to transmit copies of themselves to other computers. Windows systems always have a Startup directory for each user; any program in that folder will be executed when the user logs in. So all the worm has to do is arrange to put itself (or a shortcut to itself) in the Startup directory on a remote system. Other ways exist, some much more difficult to detect, to cause a remote computer to execute a program file that has been copied to its file system. The effects of a worm can be the same as those of a virus. Indeed, the distinction between a virus and a worm is not always clear; some malware uses both methods to spread.

Another category of malware is the Trojan horse. This is a program that apparently performs a valid functionperhaps it is a game or a supposedly "improved" version of a useful utility. But when the Trojan horse is executed some other function is performed, perhaps launching a worm or virus or performing one of the nasty things that malware does. The effects of a Trojan horse are likely to be subtle and stealthy. Unlike worms and viruses, Trojan horses are voluntarily downloaded by users, and as soon as they are recognized for what they are and the word gets out, a Trojan horse will be deleted from reputable download sites.

Another kind of malware is the logic bomb. This device is a piece of code written by one of a company's (currently employed) programmers and secretly inserted into the production operating system. As long as the programmer feeds it its daily password, it does nothing. However, if the programmer is suddenly fired and physically removed from the premises without warning, the next day the logic bomb does not get its password, so it goes off.

Going off might involve clearing the disk, erasing files at random, carefully making hard-to-detect changes to key programs, or encrypting essential files. In the latter case, the company has a tough choice about whether to call the police (which may or may not result in a conviction many months later) or to give in to this blackmail and to rehire the ex-programmer as a "consultant" for an astronomical sum to fix the problem (and hope that he does not plant new logic bombs while doing so).

Yet another form of malware is spyware. This is usually obtained by visiting a Web site. In its simplest form spyware may be nothing more than a cookie. Cookies are small files exchanged between web browsers and web servers. They have a legitimate purpose. A cookie contains some information that will allow the Web site to identify you. It is like the ticket you get when you leave a bicycle to be repaired. When you return to the shop, your half of the ticket gets matched with your bicycle (and its repair bill). Web connections are not persistent, so, for example, if you indicate an interest in buying this book when visiting an online bookstore, the bookstore asks your browser to accept a cookie. When you have finished browsing and perhaps have selected other books to buy, you click on the page where your order is finalized. At that point the web server asks your browser to return the cookies it has stored from the current session, It can use the information in these to generate the list of items you have said you want to buy.

Normally, cookies used for a purpose like this expire quickly. They are quite useful, and e-commerce depends upon them. But some Web sites use cookies for purposes that are not so benign. For instance, advertisements on Web sites are often furnished by companies other than the information provider. Advertisers pay Web site owners for this privilege. If a cookie is placed when you visit a page with information about, say, bicycle equipment, and you then go to another Web site that sells clothing, the same advertising company may provide ads on this page, and may collect cookies you obtained elsewhere. Thus you may suddenly find yourself viewing ads for special gloves or jackets especially made for cyclists. Advertisers can collect a lot of information about your interests this way; you may not want to share so much information about yourself.

What is worse, there are various ways a Web site may be able to download executable program code to your computer. Most browsers accept plug-ins to add additional function, such as displaying new kinds of files. Users often accept offers for new plugins without knowing much about what the plugin does. Or a user may willingly accept an offer to be provided with a new cursor for the desktop that looks like a dancing kitten. And a bug in a web browser may allow a remote site to install an unwanted program, perhaps after luring the user to a page that has been carefully constructed to take advantage of the vulnerability. Any time a program is accepted from another source, voluntarily or not, there is a risk it could contain code that does you harm.

Accidental Data Loss

In addition to threats caused by malicious intruders, valuable data can be lost by accident. Some of the common causes of accidental data loss are

1. Acts of God: fires, floods, earthquakes, wars, riots, or rats gnawing tapes or floppy disks.

2. Hardware or software errors: CPU malfunctions, unreadable disks or tapes, telecommunication errors, program bugs.

3. Human errors: incorrect data entry, wrong tape or disk mounted, wrong program run, lost disk or tape, or some other mistake.

Most of these can be dealt with by maintaining adequate backups, preferably far away from the original data. While protecting data against accidental loss may seem mundane compared to protecting against clever intruders, in practice, probably more damage is caused by the former than the latter.

5.4.2. Generic Security Attacks

Finding security flaws is not easy. The usual way to test a system's security is to hire a group of experts, known as tiger teams or penetration teams, to see if they can break in. Hebbard et al. (1980) tried the same thing with graduate students. In the course of the years, these penetration teams have discovered a number of areas in which systems are likely to be weak. Below we have listed some of the more common attacks that are often successful. When designing a system, be sure it can withstand attacks like these.

These and other attacks are discussed by Linde (1975). Many other sources of information on security and testing security can be found, especially on the Web. A recent Windows-oriented work is Johansson and Riley (2005).

5.4.3. Design Principles for Security

Saltzer and Schroeder (1975) have identified several general principles that can be used as a guide to designing secure systems. A brief summary of their ideas (based on experience with MULTICS) is given below.

First, the system design should be public. Assuming that the intruder will not know how the system works serves only to delude the designers.

Second, the default should be no access. Errors in which legitimate access is refused will be reported much faster than errors in which unauthorized access is allowed.

Third, check for current authority. The system should not check for permission, determine that access is permitted, and then squirrel away this information for subsequent use. Many systems check for permission when a file is opened, and not afterward. This means that a user who opens a file, and keeps it open for weeks, will continue to have access, even if the owner has long since changed the file protection.

Fourth, give each process the least privilege possible. If an editor has only the authority to access the file to be edited (specified when the editor is invoked), editors with Trojan horses will not be able to do much damage. This principle implies a fine-grained protection scheme. We will discuss such schemes later in this chapter.

Fifth, the protection mechanism should be simple, uniform, and built into the lowest layers of the system. Trying to retrofit security to an existing insecure system is nearly impossible. Security, like correctness, is not an add-on feature.

Sixth, the scheme chosen must be psychologically acceptable. If users feel that protecting their files is too much work, they just will not do it. Nevertheless, they will complain loudly if something goes wrong. Replies of the form "It is your own fault" will generally not be well received.

5.4.4. User Authentication

Many protection schemes are based on the assumption that the system knows the identity of each user. The problem of identifying users when they log in is called user authentication. Most authentication methods are based on identifying something the user knows, something the user has, or something the user is.

Passwords

The most widely used form of authentication is to require the user to type a password. Password protection is easy to understand and easy to implement. In UNIX it works like this: The login program asks the user to type his name and password. The password is immediately encrypted. The login program then reads the password file, which is a series of ASCII lines, one per user, until it finds the line containing the user's login name. If the (encrypted) password contained in this line matches the encrypted password just computed, the login is permitted, otherwise it is refused.

Password authentication is easy to defeat. One frequently reads about groups of high school, or even junior high school students who, with the aid of their trusty home computers, have broken into some top secret system owned by a large corporation or government agency. Virtually all the time the break-in consists of guessing a user name and password combination.

Although more recent studies have been made (e.g., Klein, 1990) the classic work on password security remains the one done by Morris and Thompson (1979) on UNIX systems. They compiled a list of likely passwords: first and last names, street names, city names, words from a moderate-sized dictionary (also words spelled backward), license plate numbers, and short strings of random characters.

They then encrypted each of these using the known password encryption algorithm and checked to see if any of the encrypted passwords matched entries in their list. Over 86 percent of all passwords turned up in their list.

If all passwords consisted of 7 characters chosen at random from the 95 printable ASCII characters, the search space becomes 95⁷, which is about 7x 10¹³. At 1000 encryptions per second, it would take 2000 years to build the list to check the password file against. Furthermore, the list would fill 20 million magnetic tapes. Even requiring passwords to contain at least one lowercase character, one uppercase character, and one special character, and be at least seven characters long would be a major improvement over unrestricted user-chosen passwords.

Even if it is considered politically impossible to require users to pick reasonable passwords, Morris and Thompson have described a technique that renders their own attack (encrypting a large number of passwords in advance) almost useless. Their idea is to associate an n-bit random number with each password. The random number is changed whenever the password is changed. The random number is stored in the password file in unencrypted form, so that everyone can read it. Instead of just storing the encrypted password in the password file, the password and the random number are first concatenated and then encrypted together. This encrypted result is stored in the password file.

Now consider the implications for an intruder who wants to build up a list of likely passwords, encrypt them, and save the results in a sorted file, f, so that any encrypted password can be looked up easily. If an intruder suspects that Marilyn might be a password, it is no longer sufficient just to encrypt Marilyn and put the result in f. He has to encrypt 2ⁿ strings, such as Marilyn0000, Marilyn0001, Marilyn0002, and so forth and enter all of them in f. This technique increases the size of f by 2ⁿ. UNIX uses this method with n= 12 . It is known as salting the password file. Some versions of UNIX make the password file itself unreadable but provide a program to look up entries upon request, adding just enough delay to greatly slow down any attacker.

Although this method offers protection against intruders who try to precompute a large list of encrypted passwords, it does little to protect a user David whose password is also David. One way to encourage people to pick better passwords is to have the computer offer advice. Some computers have a program that generates random easy-to-pronounce nonsense words, such as fotally, garbungy, or bipitty that can be used as passwords (preferably with some upper case and special characters thrown in).

Other computers require users to change their passwords regularly, to limit the damage done if a password leaks out. The most extreme form of this approach is the one-time password. When one-time passwords are used, the user gets a book containing a list of passwords. Each login uses the next password in the list. If an intruder ever discovers a password, it will not do him any good, since next time a different password must be used. It is suggested that the user try to avoid losing the password book.

It goes almost without saying that while a password is being typed in, the computer should not display the typed characters, to keep them from prying eyes near the terminal. What is less obvious is that passwords should never be stored in the computer in unencrypted form. Furthermore, not even the computer center management should have unencrypted copies. Keeping unencrypted passwords anywhere is looking for trouble.

A variation on the password idea is to have each new user provide a long list of questions and answers that are then stored in the computer in encrypted form. The questions should be chosen so that the user does not need to write them down. In other words, they should be things no one forgets. Typical questions are:

---

[Page 535]

1. Who is Marjolein's sister?

2. On what street was your elementary school?

3. What did Mrs. Woroboff teach?

At login, the computer asks one of them at random and checks the answer.

Another variation is challenge-response. When this is used, the user picks an algorithm when signing up as a user, for example x². When the user logs in, the computer types an argument, say 7, in which case the user types 49. The algorithm can be different in the morning and afternoon, on different days of the week, from different terminals, and so on.

Physical Identification

A completely different approach to authorization is to check to see if the user has some item, normally a plastic card with a magnetic stripe on it. The card is inserted into the terminal, which then checks to see whose card it is. This method can be combined with a password, so a user can only log in if he (1) has the card and (2) knows the password. Automated cash-dispensing machines usually work this way.

Yet another approach is to measure physical characteristics that are hard to forge. For example, a fingerprint or a voiceprint reader in the terminal could verify the user's identity. (It makes the search go faster if the user tells the computer who he is, rather than making the computer compare the given fingerprint to the entire data base.) Direct visual recognition is not yet feasible but may be one day.

Another technique is signature analysis. The user signs his name with a special pen connected to the terminal, and the computer compares it to a known specimen stored on line. Even better is not to compare the signature, but compare the pen motions made while writing it. A good forger may be able to copy the signature, but will not have a clue as to the exact order in which the strokes were made.

Finger length analysis is surprisingly practical. When this is used, each terminal has a device like the one of Fig. 5-23. The user inserts his hand into it, and the length of each of his fingers is measured and checked against the data base.

We could go on and on with more examples, but two more will help make an important point. Cats and other animals mark off their territory by urinating around its perimeter. Apparently cats can identify each other this way. Suppose that someone comes up with a tiny device capable of doing an instant urinalysis, thereby providing a foolproof identification. Each terminal could be equipped with one of these devices, along with a discreet sign reading: "For login, please deposit sample here." This might be an absolutely unbreakable system, but it would probably have a fairly serious user acceptance problem.

The same could be said of a system consisting of a thumbtack and a small spectrograph. The user would be requested to jab his thumb against the thumbtack, thus extracting a drop of blood for spectrographic analysis. The point is that any authentication scheme must be psychologically acceptable to the user community. Finger-length measurements probably will not cause any problem, but even something as nonintrusive as storing fingerprints on line may be unacceptable to many people.

Countermeasures

Computer installations that are really serious about securityand few are until the day after an intruder has broken in and done major damageoften take steps to make unauthorized entry much harder. For example, each user could be allowed to log in only from a specific terminal, and only during certain days of the week and hours of the day.

Dial-up telephone lines could be made to work as follows. Anyone can dial up and log in, but after a successful login, the system immediately breaks the connection and calls the user back at an agreed upon number. This measure means than an intruder cannot just try breaking in from any phone line; only the user's (home) phone will do. In any event, with or without call back, the system should take at least 10 seconds to check any password typed in on a dial-up line, and should increase this time after several consecutive unsuccessful login attempts, in order to reduce the rate at which intruders can try. After three failed login attempts, the line should be disconnected for 10 minutes and security personnel notified.

All logins should be recorded. When a user logs in, the system should report the time and terminal of the previous login, so he can detect possible break ins.

The next step up is laying baited traps to catch intruders. A simple scheme is to have one special login name with an easy password (e.g., login name: guest, password: guest). Whenever anyone logs in using this name, the system security specialists are immediately notified. Other traps can be easy-to-find bugs in the operating system and similar things, designed for the purpose of catching intruders in the act. Stoll (1989) has written an entertaining account of the traps he set to track down a spy who broke into a university computer in search of military secrets.