By Brijesh Balakrishnan, Vice President & Global Head of CyberSecurity Practice, Infosys
With rapid advances in technology, cyberattacks are increasingly becoming stealthier and more complex, often evading detection by security tools. We are witnessing a significant rise in APT’s or Advanced Persistent Threats, where intruders establish unauthorized access and long-term presence on targeted networks, to mine highly sensitive data. This has prompted businesses to develop robust cybersecurity strategies and observability is the latest buzzword in enterprise cybersecurity arsenal.
Observability is a thought process that provides deep insights into the data being ingested for analytics. It provides a granular view for detecting any anomalous behavior observed in the environment, based on hypotheses with deep investigation. Observability is leveraged in cybersecurity to help detect threats and unusual activities, especially low and slow attacks that are designed to evade detection by security monitoring tools. Another driving factor for the rising popularity of observability is that organizations are looking at deriving a more contextualized view, by analyzing data collected from various data sources (known as MELT – metrics, events, logs, and traces), in order to gain insights on how a system performs in terms of response time and issuing of an alert.
The need for a unified observability platform
Today organizations are drowning in structured and unstructured data, much of which remains unused. Insights from collected data can drive smarter decisions and better business outcomes. ‘The state of observability 2024’ report found that the rise of dynamic cloud-native technology stacks has unleashed a flood of data that IT Operations and security teams are struggling to sift through. It also states that 86% of technology leaders consider it impossible for teams to cost-effectively capture and analyze the data collected, as they are coping with outdated practices and fragmented monitoring tools. In addition, according to the report, the average multi-cloud environment spans 12 different platforms and services.
With organizations struggling with tool sprawl and data silos, there is a tremendous need to optimize services by consolidating the number of tools, dashboards, platforms and applications. With these tools, the shape of data varies as each tool has its own format. There is no universally accepted security data format, and this is a huge challenge. While Open Cybersecurity Schema Framework (OCSF) is emerging, its adoption and implementation on Security Datalake is still slow. By using the power of observability, security teams can gain visibility and conduct a better analysis for a comprehensive understanding of their security posture.
Streamlining cybersecurity data OCSF
To overcome the challenge of multiple tools that offer observability at multiple levels , Amazon jump-started the OCSF, an open-source data model that anyone can use and contribute to, to create analytics models and use cases and help security teams with smarter decision making.
Observability provides us with deep insights, based on analytics, and rides on the back of Machine Learning (ML) and Artificial Intelligence (AI). ML, as the name suggests learns on its own, based on its training, so models can detect suspicious activity, and generate deep insights that will complement threat detection. AI algorithms help identify patterns and deviations in AI data sets. While hallucination in AI and gen AI can result in models giving inappropriate content or wrong decisions, when clubbed with observability, ML and AI take on the role of co-pilots that ensure greater efficiency and innovation, with real-time detection of anomalies.
Advantages of observability
Observability helps security teams by providing deep insights on events of interest (EoI, uncovering patterns indicative of malicious intent. This allows them to thwart cyber intrusions, more efficiently and in a proactive manner. With cyber observability, post-incident analysis is possible, so there is no further occurrence of the same incident. A robust cyber observability solution is a strong risk management strategy in the long term.
Conclusion
While there are some privacy concerns about using observability, it is still an emerging trend that organizations are considering to bolster their security posture. As of today, with OCSF implementation, organizations are realizing that data lakes can be leveraged even further when observability is weaved into the data ingestion layer itself. By doing this, the orchestrator or schemer within the OCSF can map all data according to what is defined. This in turn, allows for forensic investigations to identify the root cause of cyber incidents, offering unparalleled insights into the threat landscape.
