Cyber risk and advisory programs that identify security gaps and build strategies to address them.
MDR that provides improved detection, 24/7 threat hunting, end-to-end coverage and most of all, complete Response.
Our team delivers the fastest response time in the industry. Threat suppression within just 4 hours of being engaged.
Be protected by the best from Day 1.
24/7 Threat Investigation and Response.
Expert hunting, research and content.
Defend brute force attacks, active intrusions and unauthorized scans.
Safeguard endpoints 24/7 by isolating and remediating threats to prevent lateral spread.
Investigation and enhanced threat detection across multi-cloud or hybrid environments.
Configuration escalations, policy and posture management.
Detects malicious insider behavior leveraging Machine Learning models.
Customer testimonials and case studies.
Stories on cyberattacks, customers, employees, and more.
Cyber incident, analyst, and thought leadership reports.
Demonstrations, seminars and presentations on cybersecurity topics.
Information and solution briefs for our services.
MITRE ATT&CK Framework, Cybersecurity Assessment, SOC Calculator & more
The Advanced Threat Analytics (ATA) team operates as eSentire’s advanced threat research and development branch. They concentrate on ways to solve the challenges posed by disparate data sets and expanding attack surfaces. Leveraging data science and machine learning expertise, the ATA team has created several proprietary and proven applications designed to identify threat actor tactics, techniques, and procedures that legacy security tools miss.
As part of our commitment to providing market-leading Managed Detection and Response (MDR) services, eSentire chose to partner with the Cyber Science Lab[1] at the University of Guelph and Mitacs[2] for two research projects. The collaboration between academia and industry aims to bring the best of both worlds to cyber security and eSentire customers.
Two promising University of Guelph students, Alex Chen and Samira Eisaloo Gharghasheh, enrolled in the Master of Cyber Security and Threat Intelligence program, were chosen to work closely with eSentire’s ATA team in an effort to research machine learning solutions to the problems that have been difficult to solve using legacy approaches.
Under the supervision of Ali Dehghantanha, director of the Cyber Science Lab and a professor at the University of Guelph, and myself, manager of the Advanced Threat Analytics team, both students spent four months analyzing large amounts of data and developing machine learning approaches to discovering adversaries and anomalous data points.
In the first project, titled “Detection of Enumeration Attacks in Cloud Environments Using Infrastructure Log Data,” Samira focused on data found in eSentire’s esLOG service.
With the complexities present in modern cloud environments, it can be a daunting task to keep track of permissions and policies. Users and service accounts often have more access than is strictly necessary, opening the doors for adversaries. Enumeration attacks are a common way for adversaries to expand their reach within a victim’s cloud environment. Once a set of credentials has been compromised or authentication tokens intercepted, the attacker will aim to discover resources they have gained access to. In order to achieve this, cloud services and accounts will be enumerated. Successful access will be further explored until a vulnerable system is found or data can be extracted.
In this project, Samira worked off AWS IAM logs that can be found and analyzed in esLOG. Through the use of Open Source red team tooling and real-world data, a wide variety of data points made up the training and validation data sets. Samira created long-short term memory (LSTM) and convolutional neural network (CNN) models to compare their performance and found that over 99% detection accuracy can be achieved.
“Classification and Anomaly Detection of Network Traffic at the Edge Using Transfer Learning” is a project Alex focused on, evaluating the benefits of transfer learning to introduce machine learning models that can accurately identify malicious network traffic in disparate customer environments.
Network traffic analysis traditionally employs large rule sets that identify well-known malicious behavior in the data stream. A subset of rules is usually specific to a customer’s environment and requires careful adjustment and monitoring when first deployed. Using machine learning, Alex attempted to reduce the need for manual intervention.
Having started out with a dataset provided by the Canadian Institute for Cybersecurity (CIC), a variety of different approaches were required to arrive at a solid model that would identify malicious actors based on network traffic. Using transfer learning techniques, a model was then trained to classify traffic captured as part of eSentire’s services. The new model used multiple fully-connected layers on top of the original one, leading to correct identification of malicious activity in over 94% of the cases.
Going forward, the Advanced Threat Analytics team, along with the eSentire organization, will continue to improve upon the above projects and integrate the detections into our portfolio. We hope to continue our partnerships and collaboration efforts in the future, and look forward to congratulating Samira and Alex on their Master’s degrees at the end of the semester.
Throughout this collaboration, the different perspectives and expertise from Samira, Alex, Ali and The University of Guelph and Mitacs have allowed all of us to grow and produce fantastic results and will be used to better serve eSentire customers around the world.
Tim leads eSentire's Advanced Threat Analytics team, working at the forefront of Machine Learning in the MDR space. He and his team deliver solutions to the most difficult to detect adversarial tactics.