Cybersecurity is an industry of espionage, crime markets, and cyber mercenaries reverse engineering the weapons of warring states. Perhaps these stories are so sensational that we’re focusing a too much on the bad guys. That’s a bold assertion in an industry driven by threat intelligence, yet our perimeters have already been invaded, and we’re occupied. Shouldn’t we focus on circling the wagons around our sensitive data?
It used to be that a perimeter breach caused a stir. Today, hacker ownage of endpoints could be another day at the office -- unless our beloved customer data has been accessed. Unfortunately, the Identity Theft Resource Center reports the number breaches stayed the same in 2015, yet hackers have been twice as successful getting to our data. Confidential data only exists on a minority of endpoints, yet cyber criminals are better at getting to these select machines than we are at defending them. We had better prioritize our defense with data-centric thinking, fast.
When I say data-centric security, I’m not really talking about the traditional role of Data Loss Prevention (DLP), which in the past has attempted to block commandeering of sensitive data. History teaches us that the perimeter is no match for today’s cybercriminals. Hackers shouldn’t be holding our crown jewels with only perimeter technologies standing between them and victory. Data-centric security begins far earlier and involves knowing where our sensitive data is supposed to be, versus where it is now.
If we know that only 10 percent of our endpoints hold sensitive data, our vital attack surface is significantly reduced. We certainly wouldn’t surrender the 90 percent, but for this vital network segment we’d have the ability to tighten policies, increase budget, and prioritize response. Our security team’s efficiency would skyrocket by focusing on 1/10th of our attack surface.
Selectively spotting data during transfer is one thing, but digging through all the sensitive data “at rest” is not that easy. Anyone who has keyword searched their hard drive knows how time and I/O intensive it is. Searching every endpoint is one thing, but proprietary document stores, intranets and cloud repositories, are difficult to access. Yet, software exists to do this, and is probably already utilized by your compliance departments for data audits and to respond to court subpoenas in the eDiscovery process.
The second problem is that knowledge of what sensitive data looks like is segregated within companies. Finding customer contact info, credit card, and social security numbers is fairly straightforward. Using jargon to locate health records and financials is a little tougher. But do you think infosec at Coca Cola knows the secret ingredients needed to spot its intellectual property? Hardly! This type of knowledge is siloed in compliance, legal, and other departments accustomed to hierarchies of privilege where certain employees know secret IP keywords, conduct searches, view intimate user data, and provide sterilized reports to the uninitiated. Herein lies the integration point between infosec and compliance: communicating this map of endpoints storing privileged data.
Recently the FCC entered a $25 million settlement with AT&T Services, related to unauthorized disclosure of 280k customer records. The uncertainty over U.S. legal penalties and new EU privacy regulations are driving a new Information Governance (IG) market. These IG folks will soon appear in new positions, such as chief information governance officer (CIGO), or other titles with the acronym IG. Their mission will be tracking, regulating, and enforcing sensitive data policies.
The growth of the IG market is sure to funnel some resources from infosec, but making data-centric security possible is a net gain. The Ponemon Institute counts the average cost of a single stolen customer record at $154, and goes as high as $363 for healthcare organizations. This means theft of only a few thousand records incurs costs in the millions.
As I’ve already mentioned, infosec efficiency could be increased with a map of endpoints housing these confidential records. Today we receive alerts as these records exit our perimeter. Why not receive a “Data Displacement Event” earlier, as data appears where it’s not supposed to be? Detecting data gathered by hackers before exfiltration could save millions.
Infosec should respond to a data displacement event with the same vigor as locating malware. Quick, put the forensic security expertise on answering the priority questions of the day! How did these records get here? Which users or programs accessed them? Have they already been transferred externally?
In the post perimeter world, it’s time for a new way of thinking. If there isn’t a boat or a shoreline nearby, nobody cares if a Tsunami hits the middle of the ocean. Similarly, the location of sensitive data means everything in judging the importance of a security incident or endpoint pwnage. It’s time we allow the people who defend an organization’s private data to benefit from those who can recognize what these records look like, and where they are.