Design of Next Generation Smart Infrastructures and Service Platforms

点击翻译按钮获取中文摘要

The convergence of three technologies -- cloud computing, software-defined networking, and machine-to-machine (M2M) communication -- provide an opportunity to create application platforms that offer unprecedented technical capabilities, scalability, energy efficiency, security, flexibility, and economics. Cloud computing provides on-demand computing power for applications at unprecedented price points. Software-defined networking allows flexible network equipment to be tailored to the needs of applications. M2M communication allows sensor and control devices to be attached to the Internet in support of new applications. In combination, these three technologies enable applications that are smart in the sense of being aware of context as well as the state of the environment. These applications will be used to manage resources in smart infrastructures (transportation, power, water, air quality, buildings etc). They will also provide novel services, such as a personal assistant that is aware of context, activity schedule, and goals of an individual.We view computing and communications resources as organized in multiple tiers, with remote massive data centers at the heart of the cloud, a smart edge providing telecom services, and then a fog consisting of gateways and sensor and actuator devices. Creating and deploying applications across this multitier cloud is challenging because of the variety of diverse technologies and incompatible protocols. Our research focuses on the design of management systems that span the cloud, smart edge, and fog to coordinate the allocation of resources to applications. Our research team brings experience designing and implementing management systems for software-defined infrastructures (SDI) that include computing and communications. Our team designed the management system for the SAVI national testbed for application platforms and also has extensive experience in the design of optimization algorithms that ensure the effective and efficient operation of the infrastructures. In this project we address four interrelated problems. First, we develop a roadmap for the deployment of computing resources in wireless and optical access networks. These resources will allow access networks to become part of an extended cloud that can be controlled by SDI managers. In particular, the access infrastructure will be virtualized so that it can be divided and shared dynamically by virtual service providers. These providers will use new business models to offer novel services that are currently infeasible or uneconomical. Our second research goal is to develop a platform for M2M and Internet-of-Things applications that uses new internetworking protocols to span the fog and the smart edge. We will also develop software-defined management to support applications end-to-end from the fog to the cloud. Our third goal is to develop graph-based optimization algorithms to support the above management systems.Our final goal involves the design of a new class of optical switches (using array waveguide gratings) that are distinguished by their ability to switch and transfer very large volumes of data. Our switches can be applied to interconnect the traffic between computers in datacenters, and can serve to collect traffic from user sites into the smart edge of the network.Our research provides benefits to Canada by generating commercializable intellectual property and by creating graduates with the skills and know-how to drive innovation.