Autonomous and proactive multi-cloud optimization using the open-source MELODIC platform (ENG)
During the last couple of years, hybrid and multi-cloud solutions are becoming very popular. With the emerging cloud options, modern enterprises increasingly rely on hybrid cloud solutions to meet their computational demands by acquiring additional resources from public clouds dynamically as per their needs. International Data Corporation (IDC), which is a leading market-research firm, in its CloudView Survey 2017, reported that 87% of cloud users have adopted a hybrid cloud strategy and 56% of the users use more than one type of cloud deployment. Still, many organizations hesitate to use Cloud computing because they have data that must stay private, and it is difficult to avoid cloud provider lock-in. Moreover, different Cloud providers offer different solutions and it could be desirable to mix and match the best offerings. The minimal deployment requirements for most cloud users are to minimize the deployment cost while maximizing the performance of their application. Cloud computing offers significant advantages over traditional cluster computing architectures including flexibility, high availability, ease of deployments, and on-demand resource allocation – all packed up in an attractive pay-as-you-go economic model for the users.
This workshop, entitled: “Autonomous and proactive multi-cloud optimization using the open-source MELODIC platform” will provide an introduction to the multi-cloud application modeling, configuration, deployment, and adaptation, including the survey of the existing Cloud Management Platforms (CMP), modeling methods, and languages.
Moreover, we will provide an overview of the latest research, the comparison of different CMPs available (such as Cyclone RightScale and Google Cloud Anthos).
During our workshop, we will use the MELODIC open source middleware platform to implement a self-adaptive deployment and reconfiguration system based on a feedback-driven control loop for an example multicomponent cloud application doing genome data mining.
Based on the MELODIC approach, we will show existing research project MORPHEMIC and its first release of Proactive and Polymorphic Adaptation, Proactive Scheduler, Self-healing Event Management System and CAMEL Designer. We will also mention existing research challenges in the area to motivate research in this direction.
Marta Różańska is a PhD student at University of Oslo and researcher developer at 7bulls.com. Marta obtained her BS and MSc degrees in Computer Science from the University of Warsaw, in 2015 and 2018, respectively. She has been working on Cloud computing optimisation for 4 years. Marta has published several research papers in this area. Her Master’s thesis investigated modelling and implementation of user preferences by utility functions in order to evaluate configuration of Cloud application deployments. Her PhD thesis is connected with utility-based optimisation of Cloud application resources. She was one of the key developers of the MELODIC platform, and now she is a Work Package leader in MORPHEMIC H2020 project. Her research interests include stochastic multi-objective optimisation, Cloud Computing, and Machine Learning.