Development of an Intent-Based Closed-Loop Security Control System for Cloud-Based Security Services
Prof. Jaehoon (Paul) Jeong at Sungkyunkwan University and Dr. Patrick Lingga who was an M.S.-Ph.D.-combined student at SKKU developed a Cloud-Based Intelligent Security Service System. They had the data models of the interfaces for this system approved as Internet Standards by the Internet Engineering Task Force (IETF) that is a De facto standards organization for the Internet.
The research group of Prof. Jaehoon (Paul) Jeong published a journal paper entitled "ICSC: Intent-Based Closed-Loop Security Control System for Cloud-Based Security Services". In this paper, they introduce the implementation of the Security Service System that supports Intent-Based Networking (IBN) intelligently addressing a user's intent, prove the concept of ICSC, and verify its performance.
When they use various security solutions together, the legacy cloud security service systems lacked the unified standardized interfaces, so an individual interface per security solution was designed and implemented to configure security policies in each vendor's security solutions and manage them. To resolve this inconvenience and inefficiency, a new Working Group called "Interface to Network Security Functions (I2NSF)" was formed in the IETF. I2NSF WG has standardized five YANG Data Models for I2NSF standard interfaces and I2NSF Applicability as Request for Comments (RFCs) that are standard documents. Prof. Jeong and Dr. Lingga contributed to this I2NSF standardization as a document editor and a YANG data model editor, respectively.
Prof. Jeong's research group have implemented and verified the ICSC System on the basis of their standardization results in the IETF I2NSF WG for the last eight years. To provide security services, this ICSC System performs two phases such as (i) Intent Fulfillment and (ii) Intent Assurance. First, in the phase of Intent Fulfillment, the intent of a user's security service request is configured in an appropriate Network Security Function (NSF) in the ICSC system. In this ICSC system, I2NSF User, which is a software used by a security administrator, composes a high-level security policy and sends it to Security Controller that is a core control and management component in the ICSC system. A Security Policy Translator (SPT) in Security Controller translates the high-level security policy into the corresponding low-level security policy that an NSF can understand. The SPT selects an appropriate NSF to be able to perform the translated low-level security policy and sends the security policy to the NSF. After receiving the security policy, the NSF performs a security service corresponding to the policy.
Second, in the phase of Intent Assurance, the ICSC system validates whether NSFs perform the requested security services well according to the user's security intent or not. The NSFs deliver their monitoring data to I2NSF Analyzer either periodically or on every occurrence of an important event. I2NSF Analyzer analyzes the NSF monitoring data by Artificial Intelligence (AI) and Machine Learning (ML) algorithms. Through this analysis, I2NSF Analyzer can find out either new security attacks or hardware issues of an NSF (e.g., the resource lack of computing power, memory capacity, and network bandwidth). For the new security attacks, I2NSF Analyzer generates Policy Reconfiguration as a low-level security policy to cope with such attacks and then sends it to Security Controller. Security Controller delivers the security policy to an appropriate NSF. Also, for the hardware issues, I2NSF Analyzer generates Feedback Information including an issue and a possible resolution and then sends it to Security Controller. Security Controller sends a request message related to the feedback information to Developer's Management System (DMS). DMS performs either the scale-up of the existing NSF or the generation of a new NSF according to the NSF hardware request message. This research measured and analyzed the metrics of security attack detection time and security attack response time for the two countermeasures such as Automatic Countermeasure with the ICSC system and Passive Countermeasure with the legacy manual system. Through the measurement of Mean Time to Detect (MTTD) and Mean Time to Respond (MTTR), it can be seen that the ICSC approach outperforms the manual approach in both the security attack detection time and the reaction time.
Currently, this research group is developing an Intent Translator that accommodates a security service request in a natural language for the sake of a security administrator in the ICSC system. This Intent Translator can translate a security intent into a high-level security policy with both Large Language Model (LLM) and Knowledge Graph (KG).
This research was performed by the Information and Communication (ICT) Standards Development Support Program of the Institute of Information & Communications Technology Planning & Evaluation (IITP) in the Ministry of Science and ICT (MSIT) of the Republic of Korea. The result of this research was published in a top international journal entitled IEEE Communications Magazine whose Impact Factor (IF) is 8.3 and that is ranked within top 5% in Journal Citation Reports (JCR).
More information:
DOI: doi.org/10.1109/MCOM.001.2400022
Provided by Sungkyunkwan University