Int2IT: An Intent-based TOSCA IT Infrastructure Management Platform
Manuel Duarte Mascarenhas, Rui Santos Cruz · 2022 17th Iberian Conference on Information Systems and Technologies (CISTI) · 2022
The introduction and widespread adoption of cloud computing has opened the door to the possibility of designing and building large scale systems with tremendous raw computing capabilities for dealing with an ever-increasing volume of data. However, the infrastructure required to support these systems became too complex for manual efforts, so practices such as Infrastructure as Code (IaC) and Development and Operations (DevOps) methodologies with programmatic orchestration and provisioning of cloud infrastructures became increasingly common. This led to the creation of the OASIS Topology and Orchestration Specification for Cloud Applications (TOSCA) standard, which has seen extensive research by academia but infrequent adoption by the industry, as, despite simplifying partly the management of infrastructures, lacks the capacity of encapsulating the desired behaviour of the system and its business-level end-goals. The system only changes its configuration through manual intervention, which is a major source of error, or through pre-defined policies that instruct actions when a specific trigger occurs. Int2IT is presented here as a proof-of-concept solution to this specific problem, as an "intent-based" infrastructure management platform that incorporates autonomic computing concepts in order to manage cloud deployments autonomously, using a TOSCA-based cloud application description. The proposed solution will be able to capture the user’s "intents", which describe the system’s end-goals, and translate them into a TOSCA-based cloud application. As a result, it can then be deployed to the cloud and autonomously managed, by utilizing a Monitor-Analyze-Plan-Execute over a shared Knowledge (MAPE-K) autonomic deployment life-cycle capable of adapting to the outside environment, ensuring that the end-goals are met to the greatest extent possible.