Policy helps to define what you consider to be valuable, and it specifies which steps should be taken to safeguard those assets.
Policy can be formulated in a number of different ways. You could write a very simple, general policy of a few pages that covers most possibilities. You could also craft a policy for different sets of assets: for example, a policy for email, a policy for personnel data, and a policy on accounting information. A third approach, taken by many large corporations, is to have a small, simple policy augmented with standards and guidelines for appropriate behavior. We'll briefly outline this latter approach, with the reader's understanding that simpler policies can be crafted; more information is given in a number of books cited in Appendix C.
Policy plays three major roles. First, it makes clear what is being protected and why. Second, it clearly states the responsibility for that protection. Third, it provides a ground on which to interpret and resolve any later conflicts that might arise. What the policy should not do is list specific threats, machines, or individuals by name?the policy should be general and change little over time. For example:
Information and information-processing facilities are a critical resource for the Big Whammix Corporation. Information should be protected commensurate with its value to Big Whammix, and consistent with applicable law. All employees share in the responsibility for the protection and supervision of information that is produced, manipulated, received, or transmitted in their departments. All employees likewise share in the responsibility for the maintenance, proper operation, and protection of all information-processing resources of Big Whammix.
Information to be protected is any information discovered, learned, derived, or handled during the course of business that is not generally known outside of Big Whammix. This includes trade secret information (ours, and that of other organizations and companies), patent disclosure information, personnel data, financial information, information about any business opportunities, and anything else that conveys an advantage to Big Whammix so long as it is not disclosed. Personal information about employees, customers, and vendors is also considered to be confidential and worth protecting.
In the course of their work, Big Whammix employees will acquire confidential information, and are responsible for protecting their own knowledge. All information stored in a tangible form at Big Whammix?on computer media, on printouts, in microfilm, on CD-ROM, on audio or videotape, on photographic media, or in any other stored, tangible form?is the responsibility of the Chief Information Honcho (CIH). Thus, Big Whammix facilities should be used only for functions related to the business of Big Whammix, as determined by the President. The CIH shall be responsible for the protection of all information and information-processing capabilities belonging to Big Whammix, whether located on company property or not. He will have authority to act commensurate with this responsibility, with the approval of the President of Big Whammix. The CIH shall formulate appropriate standards and guidelines, according to good business practice, to ensure the protection and continued operation of information processing.
In this example policy, note particularly the definition of what will be protected, who is responsible for protecting it, and who is charged with creating additional guidelines. This policy can be shown to all employees, and to outsiders to explain company policy. It should remain current no matter which operating system is in use, or who the CIH may happen to be.
Standards are intended to codify the successful practice of security in an organization. They are generally phrased in terms of "shall." Standards are generally platform-independent, and imply at least a metric to determine if they have been met. They are developed in support of policy, and change slowly over time. Standards might cover such issues as how to screen new hires, how long to keep backups, and how to test UPS systems.
For example, consider a standard for backups. It might state:
Backups shall be made of all online data and software on a regular basis. In no case will backups be done any less often than once every 72 hours of normal business operation. All backups should be kept for a period of at least six months; the first backup in January and July of each year will be kept indefinitely at an off-site, secured storage location. At least one full backup of the entire system shall be taken every other week. All backup media will meet accepted industry standards for its type, to be readable after a minimum of five years in unattended storage.
This standard does not name a particular backup mechanism or software package. It clearly states, however, what will be stored, how long it will be stored, and how often it will be made.
Consider a possible standard for authentication:
Every user account on each multiuser machine shall have only one person authorized to use it. That user will be required to authenticate his or her identity to the system using some positive proof of identity. This proof of identity can be through the use of an approved authentication token or smart card, an approved one-time password mechanism, or an approved biometric unit. Reusable passwords will not be used for primary authentication on any machine that is ever connected to a network or modem, that is portable and carried off company property, or that is used outside of a private office.
Guidelines are the "should" statements in policies. The intent of guidelines is to interpret standards for a particular environment?whether it is a software environment or a physical environment. Unlike standards, guidelines may be violated, if necessary. As the name suggests, guidelines are not usually used as standards of performance, but as ways to help guide behavior.
Here is a typical guideline for backups:
Backups on Unix-based machines should be done with the dump program. Backups should be done nightly, in single-user mode, for systems that are not in 24-hour production use. Backups for systems in 24-hour production mode should be made at the shift change closest to midnight, when the system is less loaded. All backups will be read and verified immediately after being written.
Level 0 dumps will be done for the first backup in January and July. Level 3 backups should be done on the 1st and 15th of every month. Level 5 backups should be done every Monday and Thursday night, unless a level 0 or level 3 backup is done on that day. Level 7 backups should be done every other night except on holidays.
Once per week, the administrator will pick a file at random from a backup made that week. The operator will be required to recover that file as a test of the backup procedures.
Guidelines tend to be very specific to particular architectures and even to specific machines. They also tend to change more often than do standards, to reflect changing conditions.
The role of policy (and associated standards and guidelines) is to help protect those items you (collectively) view as important. They do not need to be overly specific and complicated in most instances. Sometimes, a simple policy statement is sufficient for your environment, as in the following example:
The use and protection of this system is everyone's responsibility. Only do things you would want everyone else to do, too. Respect the privacy of other users. If you find a problem, fix it yourself or report it right away. Abide by all applicable laws concerning use of the system. Be responsible for what you do and always identify yourself. Have fun!
Other times, a more formal policy, reviewed by a law firm and various security consultants, is the way you need to go to protect your assets. Each organization will be different. We know of some organizations that have volumes of policies, standards, and guidelines for their Unix systems.
There are some key ideas to your policy formation, though, that need to be mentioned more explicitly. These are in addition to the two we mentioned at the beginning of this chapter.
Every piece of information and equipment to be protected should have an assigned "owner." The owner is the person who is responsible for the information, including its copying, destruction, backups, and other aspects of protection. This is also the person who has some authority with respect to granting access to the information.
The problem with security in many environments is that there is important information that has no clear owner. As a result, users are never sure who makes decisions about the storage of the information, or who regulates access to the information. Information (and even equipment!) sometimes disappears without anyone noticing for a long period of time because there is no "owner" to contact or monitor the situation.
People respond better to positive statements than to negative ones. Instead of building long lists of "don't do this" statements, think how to phrase the same information positively. The abbreviated policy statement above could have been written as a set of "don'ts" as follows, but consider how much better it read originally:
It's your responsibility not to allow misuse of the system. Don't do things you wouldn't want others to do, too. Don't violate the privacy of others. If you find a problem, don't keep it a secret if you can't fix it yourself. Don't violate any laws concerning use of the system. Don't try to shift responsibility for what you do to someone else and don't hide your identity. Don't have a bad time!
When writing policies, keep users in mind. They will make mistakes, and they will misunderstand. The policy should not suggest that users will be thrown to the wolves if an error occurs.
Furthermore, consider that information systems may contain information about users that they would like to keep somewhat private. This may include some email, personnel records, and job evaluations. This material should be protected, too, although you may not be able to guarantee absolute privacy. Be considerate of users' needs and feelings.
You would be wise to include standards for training and retraining of all users. Every user should have basic security awareness education, with some form of periodic refresher material (even if the refresher involves only being given a copy of this book!). Trained and educated users are less likely to fall for scams and social-engineering attacks. They are also more likely to be happy about security measures if they understand why these measures are in place.
A crucial part of any security system is giving staff time and support for additional training and education. There are always new tools, new threats, new techniques, and new information to be learned. If staff members are spending 60 hours each week chasing down phantom PC viruses and doing backups, they will not be as effective as a staff given a few weeks of training time each year. Furthermore, they are more likely to be happy with their work if they are given a chance to grow and learn on the job, and are allowed to spend evenings and weekends with their families instead of trying to catch up on installing software and making backups.
If you have responsibility for security, but have no authority to set rules or punish violators, your own role in the organization is to take the blame when something big goes wrong.
Consider the case we heard about in which a system administrator caught one of the programmers trying to break into the root account of the payroll system. Further investigation revealed that the account of the user was filled with password files taken from machines around the Net, many with cracked passwords. The administrator immediately shut down the account and made an appointment with the programmer's supervisor.
The supervisor was not supportive. She phoned the vice president of the company and demanded that the programmer get his account back?she needed his help to meet her group deadline. The system administrator was admonished for shutting down the account and was told not to do it again.
Three months later, the administrator was fired when someone broke into the payroll system he was charged with protecting. The programmer allegedly received a promotion and raise, despite an apparent ready excess of cash.
If you find yourself in a similar situation, polish up your resumé and start hunting for a new job before you're forced into a job search by circumstances you can't control.
When you write your policy, you want to be certain to include all of the various systems, networks, personnel, and information storage within your security perimeter. The perimeter defines what is "within" your control and concern. When formulating your policies, you need to be certain you include coverage of everything that is within your perimeter or that could enter your perimeter and interact with your information resources.
In earlier years, many organizations defined their IT security perimeter to be their walls and fences. Nowadays, the perimeter is less concrete.
 And may not have any concrete at all!
For example, consider the following when developing your policies:
Portable computers and PDAs can be used to access information while away from your physical location. Furthermore, they may store sensitive information, including IP addresses, phone numbers, and passwords. These systems should have minimum levels of protection, including passwords, encryption, and physical security markings. Users should have additional training and awareness about dangers of theft and eavesdropping.
Wireless networks used on the premises or otherwise connected to site resources may be connected to by outsiders using directional antennas or simply parked in a car outside the building with a laptop. Wireless networks should be configured and protected to prevent sensitive material from being observed outside, and to prevent insertion of malicious code by attackers.
Computers used at home by the organization's personnel are subject to penetration, theft, and the accidental insertion of malicious code. They may also be used contrary to organizational policy (e.g., to run a business, or host a web server with questionable content). The policy needs to make clear how these machines are to be used, protected, and audited.
Media is dense and portable. If someone makes a CD or DVD of the company financial records to use at a remote site, what happens if the media is stolen or misplaced? Policies should govern who is allowed to take media off-site, how it should be protected (including encryption), and what will happen if it is lost or stolen. They should also detail how and when previously used media will be destroyed to limit its potential exposure.
What are the policies governing people who bring their own PDAs or laptops on site for meetings or simply while visiting? What are the rules governing their connection to site networks, phone lines, printers, or other devices?
What concerns are there about shipping computers or storage devices offsite for maintenance. What if there is sensitive material on disk? What about leased equipment that is returned to the owner?
If business partners or contractors have access to your equipment, at your site or at theirs, who guards the material? How is it kept from unwanted contamination or commingling with their own sensitive data?
What policies will be in place to govern the handling of information provided to your organization under trade secret protection or license? Who is responsible for protecting the information, and where can it be kept and stored?
What policies govern non-computer information-processing equipment? For instance, what policies govern use of the printers, copiers, and fax machines? (Sensitive information on paper is no less sensitive than online information.)
Thinking about all these issues before a problem occurs helps keep the problems from occurring. Building sensible statements into your security policy helps everyone understand the concerns and to take the proper precautions.
Decide if you are going to build around the model of "Everything that is not specifically denied is permitted" or "Everything that is not specifically permitted is denied." Then be consistent in how you define everything else.
When you plan your defenses and policy, don't stop at one layer. Institute multiple, redundant, independent levels of protection. Then include auditing and monitoring to ensure that those protections are working. The chance of an attacker's evading one set of defenses is far greater than the chance of his evading three layers plus an alarm system.
Four Easy Steps to a More Secure Computer
Running a secure computer is a lot of work. If you don't have time for the full risk-assessment and cost-benefit analysis described in this chapter, we recommend that you at least follow these four easy steps:
The key to successful risk assessment is to identify all of the possible threats to your system, and to defend against those attacks which you think are realistic threats.
Simply because people are the weak link doesn't mean we should ignore other safeguards. People are unpredictable, but breaking into a dial-in modem that does not have a password is still cheaper than a bribe. So, we use technological defenses where we can, and we improve our personnel security by educating our staff and users.
We also rely on defense in depth: we apply multiple levels of defenses as backups in case some fail. For instance, we buy that second UPS system, or we put a separate lock on the computer room door even though we have a lock on the building door. These combinations can be defeated too, but we increase the effort and cost for an enemy to do that...and maybe we can convince them that doing so isn't worth the trouble. At the very least, you can hope to slow them down enough so that your monitoring and alarms will bring help before anything significant is lost or damaged.
With these limits in mind, you need to approach computer security with a thoughtfully developed set of priorities. You can't protect against every possible threat. Sometimes you should allow a problem to occur rather than prevent it, and then clean up afterwards. For instance, your efforts might be cheaper and less trouble if you let the systems go down in a power failure and then reboot than if you bought a UPS system. And some things you simply don't bother to defend against, either because they are too unlikely (e.g., an alien invasion from space), too difficult to defend against (e.g., a nuclear blast within 500 yards of your data center), or simply too catastrophic and horrible to contemplate (e.g., your management decides to switch all your Unix machines to some well-known PC operating system). The key to good management is knowing what things you will worry about, and to what degree.
Decide what you want to protect and what the costs might be to prevent certain losses versus the cost of recovering from those losses. Then make your decisions for action and security measures based on a prioritized list of the most critical needs. Be sure you include more than your computers in this analysis: don't forget that your backup tapes, your network connections, your terminals, and your documentation are all part of the system and represent potential loss. The safety of your personnel, your corporate site, and your reputation are also very important and should be included in your plans.