Keynote: Enhancing Distributed Computing with Programmable and Open Optical Networks
Data Analytics, Compression, and Management
Performance/Productivity Measurement and Evaluation
TimeThursday, 12 November 20202:30pm - 3:15pm EDT
DescriptionEdge computing is an attractive solution to efficiently provide compute resources to many applications that demand specific QoS requirements. The edge computing resources are in close geographical proximity to where the applications’ data originate from and/or are being supplied to, thus avoiding unnecessary back and forth data transmission with servers and data centers located far away. In this talk we describe a federated edge computing system in which compute resources at multiple edge sites (cloudlets) are dynamically aggregated together to form virtual super-cloudlets and best respond to varying application-driven loads. In its simplest form a super-cloudlet consists of compute resources available at two cloudlets that are (temporarily) interconnected by dedicated optical circuits deployed to achieve low-latency and high-speed data exchanges. The metro area optical transport network must be programmable in order to provision these optical circuits in a timely fashion. In addition, an open optical network deployment would in this scenario offer a significant advantage to the cloudlet owners (which are willing to share their cloudlet resources to create super-cloudlets) in that every owner would be free to independently choose the supplier for their optical equipment deployment. We report on our initial experimental testing of the super-cloudlet architecture using the largest public OpenROADM demonstration up to date (comprising Dense Wavelength Division Multiplexing equipment from six major suppliers) and describe how the Programmable Optical network (PROnet) Orchestrator handles the resources available in the open optical metro network and connected cloudlets to achieve three key functionalities, i.e., service placement, auto-scaling, and offloading.