Senior Cloud and Big Data Architect, Samsung R&D institute
Dariusz Król joined Samsung R&D institute in Poland, Krakow in 2016 as a Senior Cloud and Big Data Architect. His current activities involves development of Database as a Service solution based on Openstack Trove, while in recent past he was also involved in various projects oriented towards analytical use cases. He is a Ph.D. in computer science since 2014 with specialization in scalability, data management and cloud and grid technologies. In 2014-2015, he was a postdoc at University of Southern California, Information Sciences Institute, where he participated in the PANORAMA (Predictive Modeling and Diagnostic Monitoring of Extreme Science Workflows) project, where he was responsible for performance monitoring and analysis of large-scale scientific workflows. Before joining Samsung Electronics, he was an assistant professor at AGH University of Science and Technology, Krakow, Poland and a specialist at ACC Cyfronet AGH, which is the biggest academic computer center in Poland. He was involved in a number of international and national scientific projects regarding large scale computational infrastructure, data analysis, resource management and parallel computing. He is co-author of over 50 scientific articles published in peer-reviewed international journals and conferences proceedings.
Platform Intelligence solutions for OpenStack
Cloud native applications intend to provide much more flexibility than legacy application by operating on an elastic infrastructure and supporting additional features like scaling, alerting and failover. However, implementing these features based on manual operations is error prone and require administration effort. To make them automatic, and to follow Autonomic Computing principles including as Self-Defining, Context-Aware, Self-Configuring, Self-Healing, Self-Optimizing, and Self-Protecting, additional services have to be integrated into the platform. During presentation, we will focus on several use cases including predictive auto-scaling, storage space prediction and noisy neighbor detection will be discussed in relate to machine learning models oriented towards operational intelligence. In additional, Openstack services like Senlin, Aodh and Vitrage will be compared.