Multicloud management platform – single gate to multicloud world
MELODIC allows for implementing applications that support various types of components: virtual machines, containers, big data (Apache Spark) platforms and serverless components. MELODIC is integrated with leading cloud service providers: AWS , Azure and Google Cloud Platform , as well as with service providers using OpenStack. Integration with local cloud providers is also being carried out.
In addition, unlike Kubernetes, which only allows you to optimize resources within an existing cluster, MELODIC allows you to dynamically add cloud resources as the application needs, as well as delete them when they are not needed.As a result, the use of resources is tailored to the needs of the application and there is no need to maintain an oversized infrastructure
Multicloud applications modelling
The key element of the platform is based on the TOSCA standard (Topology and Orchestration Specification for Cloud Applications) the CAMEL language (Cloud Application Modeling and Execution Language), allowing to describe app requirements and infrastructure independently of a specific supplier, as well as selecting the most optimal implementation model depending on the characteristics of the application. An additional element of the platform is the ability to optimize Big Data solutions and data locality Awareness.
The process of modeling/describing applications in CAMEL language within the Melodic platform first includes defining its components, connections between them, as well as requirements regarding the performance and resources, along with the way of implementing the application. In the next step, the implementation configuration is automatically optimized – the platform “decides” which and where the infrastructure should be used.
Cloud computing resources optimization
Initial optimization is made on the basis of parameters specified by the user. Optimization is one of the strongest points of the platform. Advanced methods based on Constraint Programming and Reinforcement Learning (Stochastic Learning Automata) are used to solve the optimization problems. MELODIC also includes a unique module for assessing the usability of a business implementation, based on an adaptive usability function. On the basis of a specific, optimal configuration, a precisely defined infrastructure is automatically created for selected cloud service providers (virtual machines with set parameters) and then the application is implemented along with connection settings between components. After implementation, the application is monitored – collected are, among others, metrics defining the characteristics of its operation, which at the same time constitute the basis for automatic optimization of application implementation, basing on the current values of the metrics.