In the world of Information Security (infosec), there are two main philosophies: compliance-based infosec and threat-based infosec. Compliance-based infosec means meeting a set of written security standards designed to fulfill some goal such as the requirements of statute law or financial information privacy requirements. Threat-based infosec, on the other hand, means applying information security controls in reaction to (or anticipation of) threats that organizations currently (or soon will) face.
Compliance-based infosec is generally applied smoothly across the organization. In other words, all the security controls mandated in the security standard must be put in place by the organization, and the relative effectiveness of each control is largely ignored. In contrast, security controls are applied in a hierarchical manner in threat-based infosec. The most effective or greatly needed security controls are applied first according to the threats that are most likely to occur or that will cause the most damage to the organization if they do occur.
The difference is sort of like the defensive strategy of the Chinese versus that of the Normans in post-conquest England. The Chinese built very long walls that went from one end of their territory to the other. Their goal was to keep out all invaders everywhere. This is a grand idea, but takes a very large amount of resources to implement and maintain. In practice, it takes tons of men and infrastructure and the defensive capabilities at any one place are spread thin. The Normans in England, on the other hand, built strong castles with many layers of defense in strategic locations where the threats were greatest and where it was easiest to support neighboring castles. In practice, there are fewer defenses at any one point, but the places where defenses are implemented are very strong indeed. Both of these strategies have merit, and are really driven by the particular set of circumstances faced by the defender. But which is better for your organization? Let’s look at compliance-based infosec first.
Compliance-based infosec, when implemented correctly, is really the best kind of defense there is. The problem is, the only place I’ve ever seen it really done right is in the military. In military information security, failure to protect private information can lead to death and disaster. Because of this, no expense or inconvenience is spared when protecting this information. Everything is compartmentalized and access is strictly based on need to know. Every system and connection is monitored, and there are people watching your every move. There are rules and checklists for everything and failure to comply is severely punished. In addition, finding better ways to protect information are sought after, and those that come up with valuable ideas are generously rewarded.
This is not the way compliance-base infosec works in the private sector, or even in non-military government agencies. First, statute law is tremendously vague when discussing implementing information security. Laws make broad statements such as “personal health information will be protected from unauthorized access or modification”. Fine. So a group of people get together and write up a body of regulations to further spell out the requirements organizations need to meet to comply with the law. Unfortunately, you are still dealing with pretty broad brush strokes here. To try to get a handle on things, agencies and auditors rely on information security standards and guidelines such as are documented in NIST or ISO. From these, baseline standards and requirements are set down. The problems here are many. First, baseline standards are minimums. They are not saying “it’s best if you do this”, they are saying “you will at least do this”. However, typical organizations, (which generally have very limited infosec budgets), take these baseline standards as goals to be strived for, not starting points. They very rarely meet baseline standards, let alone exceed them. Also, NIST and ISO standards are not very timely. The standards are only updated occasionally, and they are not very useful for countering new and rapidly developing threats. So, unless your organization is really serious about information security and has the money and manpower to make it work, I would say compliance-based infosec is not for you. I know that many organizations (such as health care and financial institutions) are required to meet baseline standards, but remember what happened to Target last year. They were found to be compliant with the PCI DSS, but still had tens of millions of financial records compromised.
Now let’s look at threat-based infosec. To implement a threat-based information security program, the organization first looks at the information assets they need to protect, the threats and vulnerabilities that menace them and the consequences that will ensue if those information assets are actually compromised (basic asset inventory and risk assessment). They then prioritize the risks they face and decide how to implement security controls in the most effective and efficient way to counter those particular risks. That might mean implementing strong egress filtering and log monitoring as opposed to buying the fanciest firewall. Or it might mean doing something simple like ensuring that system admins use separate access credentials for simple network access and administrative access to the system. Whatever controls are applied, they are chosen to solve particular problems, not to meet some broad baseline that is designed to meet generally defined problems. Also, threat-based infosec programs are much better at anticipating and preparing for emerging threats, since reassessments of the security program are made whenever there are significant changes in the system or threat picture.
These are the reasons that I think most of us in non-military organizations should go with threat-based infosec programs. Even those organizations that must meet regulatory requirements can ensure that they are spending the bulk of their infosec money and effort on the effective controls, and are minimizing efforts spent on those controls that don’t directly counter real-world threats. After all, the laws and regulations themselves are pretty vague. What counts in the long run is real information security, not blind compliance with inadequate and antiquated baselines.
Thanks to John Davis for this post.