RECAP Workload Predictor for Globally Distributed Services
The RECAP Workload Predictor for Globally Distributed Services location-aware workload prediction tool for Web-hosted services is based on the predicted number and location of users and their origin. The predictor makes suggestions about the total amount of compute resources needed to satisfy the load and suggests where these resources should be placed worldwide to minimise latency for the users. The RECAP Workload Predictor for Globally Distributed Services has integration points with other RECAP IP, particularly the Monitoring and Storage Infrastructure and the Application Configuration and Deployment Optimiser.
Benefits: The RECAP Workload Predictor for Globally Distributed Services enables organisations to optimise resource requirements and utilisation and thus improve performance for their service back-end whilst, at the same time, minimising provisioning costs. It transparently allows non-data scientists to employ machine learning algorithms to perform configurable predictions of the load on their system and obtain a forecast of the global origin of servers and resources. In turn, this enables proactive adaptation of distributed service provisioning to load changes further reducing overprovisioning.
Contact: IMDEA Networks Institute, Spain | Rafael García Leiva | firstname.lastname@example.org