Autonomous Resource Allocation for Edge Infrastructures

The optimization of resource allocation in cloud computing environments is a crucial problem with particular research interest and direct application to a multitude of commercial applications. A prominent class of applications that strongly emphasize this problem pertain to the Network Function Virtualization (NFV) paradigm, which promotes the implementation of typical network functions, such as firewalls, load balancers, and intrusion detection systems, in the form of virtual machines or containers. This obviates the need for hardware-based network function deployments, effectively rendering NFV-enabled networks more scalable, reliable, and cost-effective. However, to fill up its promise, NFV shall incorporate sophisticated mechanisms that can promptly compute optimized resource allocations, subject to volatile user demands and energy consumption constraints. Along these lines, the main objective of ARCADIA is to investigate, design, and evaluate innovative computational frameworks and methods for optimized resource allocation in cloud computing environments. Concretely, ARCADIA focuses on a) systems exhibiting dynamic workload characteristics, and b) environments with high energy consumption requirements due to simultaneous and continuous operation of computer clusters and equipment. Both of these features can be found in edge systems, and specifically in edge data centers, which have become a pivotal computing part of next-generation networks.