Distributed Algorithms for Spatial Retrieval Queries in Geospatial Analysis

Abstract

Cloud infrastructures needs to become more smarter and dynamic in managing resources, requiring new solutions and tools to cope with processes of configuration and reconfiguration, automatic scaling, elastic computing and healthiness control. This paper presents a Smart Cloud solution for automatically managing cloud infrastructures, and programming the monitoring aspects directly from the configuration and deploy phases. This means that the proposed solution can enforce smart cloud intelligence to almost all Cloud Configuration management systems and orchestrators by using REST calls and XML files. The solution proposed includes a Smart Cloud Engine, an advanced Supervisor and Monitor, and a Knowledge Base. The Knowledge Base is grounded on a cloud ontology modeling cloud resources and relationships, Service Level Agreements and their evolution, constraints, metrics for assessment, etc. The Knowledge Base enables the reasoning on cloud structures and permits to implement strategies of efficient smart cloud management and intelligence: (i) verify and validate the configurations and related changes according to the cloud model and available resources; (ii) easily generate the processing control for verifying cloud resource healthiness and service level agreement verification; (iii) take decision about dynamic reconfiguration on cloud, for example for scaling, balancing, migration, etc. The proposed solution has been validated in the context of ICARO project on complex cloud configurations with good performance and low operating workload. The validation has also assessed the reliability and the scalability.