New approaches related with Big Data, peta- and exa-scale computing are going to dramatically change the design, development and exploitation of highly demanding applications, such as the HPC ones. The growing requirements of modern data- and compute-intensive applications ask for new models for their development, deployment and execution. Nowadays we are noticing important changes in the way High Performance Computing (HPC) providers are dealing with the demand. Through the paper, we detail the workflow of the proposed OaaS, illustrate its usage with several use-cases and discuss its applicability in different scenarios. server role) while potentially providing high-performance computing (HPC) capabilities for their execution.
Thus, third-party players can focus on the development of optimization algorithms (i.e. client role) may request the resolution of a network resource allocation problem through the execution of an algorithmic technique. The proposed framework is based on a classical client-server architecture, and through a REST-based API, holders of network infrastructures (i.e. In this paper, we present a network OaaS framework built as an extension of the Net2Plan open-source network planning tool. Automated network optimization procedures on top of network controllers promote a client layer of third-party applications that enable optimization-as-a-service (OaaS) business opportunities. In this context, network optimization is essential to efficiently employ network resources, relying whenever possible on automated procedures and exploiting open application programmatic interfaces (APIs) inherent of SDN/NFV ecosystems. At the same time, the advent of Software-Defined Networking (SDN) and Network Function Virtualization (NFV) technologies create transport ecosystems with unprecedented network control and resource dynamicity capabilities. Carrier-grade telecommunication infrastructure must support an ever-increasing traffic volume and dynamicity in operationally-complex networks characterized by multiple domains, multiple technologies and equipment from multiple vendors.
0 kommentar(er)
