The scientific paper entitled “PrEstoCloud: A Novel Framework for Data-Intensive Multi-Cloud, Fog, and Edge Function-as-a-Service Applications”, co-authored by UBITECH, has been accepted and published online at the Information Resources Management Journal (IRMJ) of IGI Global. Giannis Ledakis, Head of UBITECH’s Computing Systems, Software and Services research group, and his co-authors from NTUA and JSI review prominent fog computing frameworks and discuss some of the challenges and requirements of Function-as-a-Service (FaaS) -enabled applications. Moreover, they propose a novel framework able to dynamically manage multi-cloud, fog, and edge resources and to deploy data-intensive applications developed using the FaaS paradigm. The proposed framework leverages the FaaS paradigm in a way that improves the average service response time of data-intensive applications by a factor of three regardless of the underlying multi-cloud, fog, and edge resource infrastructure.
In particular, The proposed framework allows not only the execution of microservices with cloud, fog, and edge devices but also decides, at run-time, the shifting of processing tasks from edge devices to multi-cloud or fog resources and vice-versa. To facilitate this, proposed framework allows DevOps to define constraints and execution preferences about microservices that can be executed either on cloud, fog, or edge resources. The proposed framework decides during run-time to offload/onload processing tasks on specific types or instances of edge resources to/from multi-cloud resources. Such decisions are based on the current context of the resources used, the microservices status with respect to QoS and the present and estimated workload.
The framework that was developed provides the capability to run functions by using microservices, specifically, with the deployment of Docker images. Therefore, the framework supports the deployment of applications built using the microservice and FaaS paradigms through the use of lambda functions, as well as the deployment of any applications capable of exploiting the distributed deployment paradigm. Deployment blueprints are expressed in an extended topology and orchestration specification for cloud applications specification (TOSCA) that considers edge and FaaS-related constructs. These blueprints are exploited in processing workflows that enable the deployment of microservices.
