Getting threat intelligence right
Threat intelligence feeds provide valuable information to help identify incidents quickly, but only if they are part of an intelligence-driven security programme.
When shoppers went about their usual routine at two malls in the heart of Singapore on 20 May, they were greeted with an unexpected message that popped up on the stores’ digital directories.
The message came from the cyber criminals behind the unprecedented WannaCry ransomware attack, demanding $300 worth of bitcoin payments to unlock the files on systems that powered the store directories.
Although the local supplier of the malls’ digital directories fixed the problem promptly, such ransomware attacks, which could embarrass businesses and cause inconvenience for customers, could be averted with good cyber hygiene – and some help from threat intelligence feeds.
In the wake of the WannaCry attack, threat intelligence feeds are essential to alert security analysts to threats that are proliferating but have yet to infiltrate company networks, says Charles Lim, industry principal at Frost and Sullivan’s cyber security practice in the Asia-Pacific (APAC) region.
Threat intelligence feeds may include data such as IP blacklists, malware hashes and signatures, malicious and phishing URLs, vulnerability lists, indicators of compromise (IOCs) such as IP addresses that suggest a breach has occurred, and command and control (C2) domains that are used to orchestrate attacks.
These data points are typically consolidated by security information and event management (SIEM) and intrusion monitoring systems, and could provide insights into the possible identities of attackers, the methods they are using and the systems they are targeting.
This is especially useful in situations such as the WannaCry attack, where security teams do not always have the time and resources to constantly identify emerging threats and implement appropriate safeguards, says Peter Sparkes, senior director of cyber security services at Symantec Asia Pacific and Japan.
According to Symantec, the biggest users of intelligence feeds in APAC are Australia, Singapore and Japan, mostly organisations in the financial services and government sectors.
That should come as no surprise given that finance, insurance and real estate are among the most targeted industries, according to Symantec’s latest Internet Security Threat Report.
While threat intelligence feeds provide valuable information to help identify incidents quickly across an enterprise, they are generally based on known, observed information.
Much of today’s threat intelligence is supplied as IOCs – essentially fingerprints of known attacks or attackers, says Kane Lightowler, managing director of Carbon Black in Asia Pacific and Japan. “IOCs may provide great value against previously observed attacks, but offer limited insights on new attacks and attack methodologies.”
Sparkes agrees, noting that intelligence feeds require a “patient zero” – the first organisation or person to see the attack and record the IOCs before others can benefit from it.
Lightowler says patterns of attack are more effective against both known and unknown threats because they focus on the actual behaviour and techniques of the attacker, rather than fingerprints. Regardless of the type of threat, intelligence can be very valuable if it is actionable in near real time.
“Threat intelligence is time-sensitive, so the value of information depreciates once it is published,” says George Lee, senior director at RSA Asia Pacific and Japan. “The flip side is, you never know whether the bad guys also subscribe to the same threat intelligence as you do, so they may change their tools, techniques and tactics to stay below the radar.”
To improve the efficacy of intelligence feeds and detect new threats, threat intelligence suppliers rely on a variety of sources. For example, FireEye not only gathers threat-related data from sensors on the internet, but also collects threat information from incident response teams at its Mandiant subsidiary that works with companies to investigate security breaches.
“We can see the tactics, techniques and procedures undertaken by the bad guys once they get on target,” says Tim Wellsmore, director of threat intelligence at FireEye. “We also have teams that engage with the bad guys on the darknet, and we will overlay all that intelligence with what we’ve been getting through our sensors and incident response teams.”
Context and integration critical
Organisations are more likely to get the most out of intelligence feeds if they can establish context around the data they are receiving.
“Feeds should not be something that you line up and fire into an organisation,” says Wellsmore. “People who did that in the past became overloaded with data and still did not know what their threat environment really looked like.”
In fact, Wellsmore contends that intelligence feeds are really just data feeds, and that intelligence arises only when an organisation can ascertain threats specific to its operating environment.
While traditional definitions of threat intelligence are limited to classic network and IT security, a broader, more modern perspective looks at an organisation’s total attack surface and all areas of risk to the business, says James Carnall, vice-president of the cyber security centre at LookingGlass.
That means taking into account the organisation’s industry, location, software and network particulars, vulnerabilities, physical threats to employees and executives, as well as reputational and political risks, brands and trademarks, and customer goodwill.
At the end of the day, an actionable intelligence feed needs to lead to some action, whether it is a change in security posture, handling of traffic, updates to policies, reduction in risk, countering an attack or other measurable, reportable business value.
“Data with analysis and context becomes actionable information, and once it is actually able to be acted upon, only then does it become intelligence,” says Carnall.
According to Carnall, the integration, interpretation and implementation of a complete intelligence-driven security programme is key to giving threat intelligence value.
But this is not easy to achieve, because many organisations face constraints on budget, tools, knowledge and specialised talent, which are in high demand and short supply.
Sim Beng Hai, technical sales manager for Eset Asia Pacific, advises organisations to standardise their threat intelligence on Stix, an XML format for conveying data about cyber security threats in a common language that can be easily understood by humans and security technologies.
Finding the right threat intelligence supplier
With threat intelligence being strategic to an organisation’s cyber security practices, James Carnall, vice-president of the cyber security centre at LookingGlass, offers some tips on what to look out for when choosing a threat intelligence supplier:
Consider the supplier, not just the data. There is a lot of startup activity and some cool things bubbling. However, organisations must not build critical enterprise security processes around data feeds or suppliers that may not survive long enough to justify the investment in integration and training. Suppliers with long histories, credible references and mature security themselves should be a key part of the evaluation.
Consider integration path as a key element of the buying process. Don’t buy before you know how you will integrate and operationalise. LookingGlass has seen organisations waste months of subscription costs because they buy feeds without a clear idea of where and how to integrate them. Better still, deal with a supplier that has pre-built integration tools/scripts, provides services to do it for you or, best of all, provides the feeds, software and devices to operationalise the intelligence.
Be very wary of noisy, high-volume but low-fidelity or low-specificity indicators, such as blacklists that are at the IP level when full-path individual URL level is what is required. From “alert fatigue” to constant blocking of legitimate sites, few things will kill a threat intelligence programme more quickly than blocking the CEO’s access to CNN because of one malicious ad link, or because one bad site is on a shared IP with 5,000 other perfectly good websites.
Standards aside, Carnall says organisations should consider the goals they want to achieve when integrating threat intelligence into security processes.
For example, if the goal is to defend a network, then high-fidelity indicators, clear risk-scoring schemes and orchestration software could be used to set dynamic blocking or traffic rules on the network, in as close to real time as possible. “This is where IPs, domains and full-path URLs go into your gateways and the like to mitigate tactical threats,” he says.
When done right, threat intelligence can throw attackers off course. “If proper orchestration is also in place, mitigation enabled in the network fabric can then block, sinkhole, or even modify the traffic being sent back to the attacker to stop data loss before it happens or even mislead the attacker through intentional misinformation,” says Carnall.