Tuesday, April 14, 2015

March Newsletter - Big Data Analytics and Cyber-Security

Cyber security analytics is rapidly becoming a Big Data application for one simple reason, large organizations are collecting, processing and analyzing more and more data in order to effectively address the new cyber threat landscape.
Big Data Analytics and Cyber-Security is Vega’s topic of the month.
You can find our monthly technical review at page 3, and few examples of Israeli solutions for this topic at page 5.



A CHANGING THREAT LANDSCAPE
The cyber threat landscape has changed dramatically over the last 5 years. The new industrialization and internationalization of digital criminality combined with the limited legal responses available have enabled the dramatic growth and convergence of both simple and
sophisticated attacks.


It is now generally accepted, not only in the security profession but also in the Boardroom, that every organization will be attacked in one form or another on a regular basis. Some of those attacks will inevitably succeed. This is driving changes in the defensive stance of leading firms, where a greater emphasis is now being placed on identifying and limiting damage from successful attacks. Previously many firms had the stated goal of preventing all possible attacks.
As a direct result, effective and efficient security operations have become a key cyber defense capability within many leading organizations. Innovative and leading commercial organizations are now building increased security monitoring and security analytics capabilities to sit alongside effective threat intelligence and critical incident management capabilities. Their goal is to predict, to limit and to manage the inevitable attacks they will face.

INCREASED MONITORING LEADS TO BIG DATA
Cyber security analytics is rapidly becoming a Big Data application for one simple reason: large organizations are collecting, processing and analyzing more and more data in order to effectively address the new cyber threat landscape.
The promise of Security Information & Event Management (SIEM) technologies was to deliver advanced analytics capabilities. The reality is that SIEM products weren’t
designed for Big Data analytics and generally cannot meet the rapidly evolving needs that leading commercial organizations now demand.
SIEM does provide a good foundation for security monitoring in providing a near real-time signature or rules-based detection capability to look for known threats. SIEM is also great for compliance and reporting. However, SIEM does not scale to detect the unknown threats across all the available data. Data often has to be pre-filtered before being loaded in to a SIEM. This effectively presupposes where the risk lies. SIEM cannot do the advanced security analytics that are required today.
It is likely the SIEM platforms and the current range of Big Data cyber security analytics platforms will move towards convergence over the next five years. However, this is new ground for both groups of vendors and for now separate products remain necessary to achieve the full potential benefits of each.

BEHAVIORAL ANALYTICS FOR DETECTION
Behavioral analytics understand past human behavior, predict future behavior and identify anomalous behavior. Behavioral analytics have been used extensively in fraud detection and prevention because different individuals naturally display different behaviors and legitimate behaviour is practically always different from that exhibited by a fraudster.

Behavioral analytics takes advantage of this fact. Rather  than just looking for specific indicators, behavioral analytics combines knowledge with monitoring to determine if behavior is expected and legitimate, or suspicious.
Behavioral analytics is a Big Data challenge not only because of the volumes of data involved, but also because of the need to bring a wide variety of data sources and formats together to create a full picture. Cyber security analytics is increasingly adopting behavioral Analytics, from the fraud detection field, in order to address the reality that traditional security solutions have proven ineffective against the incredible variety and volume of digital criminality, such as cyber espionage, cyber crime, hacktivism and the insider threat. This emerging cyber and fraud threat detection convergence means that the benefits realized from behavioral analytics in combating both will increasingly drive even greater operational efficiencies and investment decisions.
Behavioral analytics are proving to be more robust, enduring and effective than traditional signature and rules based analytics. Figure 1  demonstrates the contrast between the two. In short, an organization using behavioral analytics will find anomalies that other point solutions and systems cannot.

A BIG DATA PLATFORM FOR INVESTIGATION
The operational cost of the technology and people required to effectively detect, triage and investigate security incidents is too high. Limits on data collection, non-interoperable tooling and subsequent data mining means that once suspicious indicators have been identified it can take weeks for a cyber-investigator to collect and analyze the data from across a large enterprise required to identify the appropriate response.
Big Data platforms can enable faster query times and a more seamless approach for security analysts retrieving and analyzing data across multiple sources and formats.
In most cases Big Data Cyber security solution will comprises of three core components:

  • Platform - Massively scalable technology platform that correlates data acquired from across the IT infrastructure
  • Analytics - Behavior-based threat detection using unique attack models 
  • Investigator - Powerful threat intelligence management and investigation toolset providing visualizations, rich contextualization and correlation of threats, indicators, events and alerts.


SUMMARY
As a result of the rapidly changing and expanding cyber threat landscape many organizations have increased their security monitoring capabilities both in terms of volume and variety of data. As such Cyber Security analysis is now becoming a Big Data problem for both the detection and investigation of incidents.