Anybody who was at the RSA Conference a few weeks ago can attest to the fact that "threat intelligence" has become the catchphrase du jour. As was the case with "cloud" a few years ago and "data analytics" last year, every vendor on the show floor was hawking a product or service which claimed to increase my understanding of the bad guy and thus improve my ability to better defend my enterprise.
As I began to dig into these products and query my fellow practitioners, I became concerned about what the vendors were calling intelligence and the real value this latest array of tools and services would (or would not) provide. Many of my discussions quickly focusd on the concepts of intelligence and intelligence collection and how to apply these to concepts practically. While it's been years since I worked in Army Intelligence, some lessons certainly spill over into the corporate world. Allow me to share some of my thoughts on these concepts for your consideration...
Intelligence Defined. With the term threat intelligence being widely (overly?) used these days, it's worth stepping back and understanding the difference between data, information, and intelligence. Data are facts; they are immutable and unchangeable. Information, simply put, is data in context. Intelligence is information that is extracted and revealed from an analysis of given data and information.
Confused? Let me give you an example. Take the number 3015178088. This number is piece of data, devoid of any context. We can attempt to provide some context to this data in an attempt to provide some level of information to the reader. Examples:
3,015,178,088 -- a number in excess of three billion
30151-70808 -- an overseas telephone number, most likely European
(301) 517-8088 -- a North American phone number
As you can see, adding context to the data provides different information to the consumer.
In this particular example, the North American telephone number is the correct context. Now let's provide additional data abd information to the reader:
301 is one of the area codes for Maryland
I lived in Maryland from 1995 to 2003
Given these pieces of data and information, you might be able to extract the intelligence that the phone number listed was one of my old phone numbers. That piece of information is not stated anywhere within the provided data/information, but through simple analysis and some deductive reasoning, it can be extracted with a reasonably high level of confidence. Consumer organizations regularly extract "business intelligence" from data collected with member loyalty programs in an attempt to focus its marketing and sales efforts with laser-like precision -- occasionally to the point of becoming quite intrusive.
Actionable Intelligence Collection. In order for intelligence to be actionable (i.e., result in the abiliity for me to do something better/smarter/differently/more effectively than I am at the moment), the intelligence collection effort within the enterprise must be organized and deliberate. Intelligence collection is more than information gathering/information consumption; it requires the enterprise to remain focused on fanatically answering three seemingly simple questions:
What do I need to know? What are the most important questions for you to get answered? In the military we referred to these questions as priority intelligence requirements (PIRs). Intelligence collection efforts should be focused on answering these questions first and foremost. Note that determining these questions may be simpler than you think. I remember an intelligence exercise from my GI days involving the transport of relief supplies into a fictional European country via military convoy. As the exercise assumed a hostile force which occasionally disrupted transports along the one major highway into the area, the #1 PIR each was always "Is the road open for travel?" I would imagine that some of the PIRs for most enterprises would be equally straightforward. Some examples:
Are bad guys in my environment right now?
Is sensitive data leaving my environment in an unauthorized fashion?
Which bad guys trying to get into my enterprise?
Where are the most likely/most vulnerable attack points?
What is the best way to get the answers I need? Folks, PIRs can (and should) be answered by a multitude of sources. These include (but are not limited to)
News reports
Existing enterprise tools
Communications via professional organizations
Organizations which monitor threat activity regularly (CERTs, ISACs)
While a "threat intelligence" platform or service, properly constructed, might provide indications and warning about an imminent attack it may be argued that existing sources of data from within the enterprise are better suited to determining the current state of attack if properly monitored and utilized. Indeed, focused monitoring and analysis of open-source information providers may provide reasonably accurate and timely indications and warning of threats and attacks against the enterprise.
What do I intend to do with the intelligence gathered? Intelligence collection should not be an academic exercise. Answering your PIRs should drive action within your environment. If fulfilling a PIR does not drive even a minimal course correction on the actions and activities of the enterprise, then you need to consider whether or not you are answering the right questions...or whether or not you need to adjust you efforts down to that which is actionable within your current culture. This last phrase may seem like an anathema to the security professional, but given limited resources we must constantly balance our collection efforts against our execution priorities lest security become simply an academic exercise.
* * * * *
Given the aforementioned definitions, many of the "threat intelligence" products and services being advertised today provide threat information versus threat intelligence -- i.e., they are providing yet another credible data/information source to the enterprise defenders versus true intelligence. While this data source may be more focused and more useful than other data sources available, without an understanding of organizational PIRs this data source becomes yet another firehose of "stuff" which the security team must consume lest it drown. Worse, without an ability to guide and focus collection efforts across the enterprise the security team may be looking for indications & warning regarding attacks from largely ineffective (and probably expeensive) sources.
Threat "intelligence" products and services do have a place within the enterprise...but only if the enterprise is prepared to utilize and absorb the information provided in a cogent and thoughtful manner. While the temptation to garner information from closer to the bad guys is appealling, having yet another firehose to drink from will only hasten the drowning if you're not careful.
My two cents...