Smart city platforms on multitier software-defined infrastructure cloud computing

We identify requirements and then propose a multi-layer architecture for smart city platforms (SCPs) based on an infrastructure layer operating as a multi-tier computing cloud using software-defined infrastructure (SDI), a platform layer providing data dissemination, and a Business-Intelligence-as-a-Service (BIaaS) layer offering analytics services to support smart city applications through open APIs. The multitier cloud enables the end-to-end orchestration of resources from sensor/actuator to smart edge and to core massive datacenter. The SDI provides the necessary heterogeneous resources, including computing, storage, GPUs, programmable hardware, sensors and networking, that are all managed jointly within a single domain or across multiple domains through federation. The data dissemination services allow individuals, organizations, and public agencies to share real-time or historical data within a domain or across domains. The analytics services provide intelligence by acting on data in support of decision-making. Public and private application providers use these services to create smart city applications. We discuss our experience implementing and operating a Canada-wide multitier SDI cloud and a Greater Toronto Area (GTA) platform for smart transportation, and we present extensions of these platforms to handle smart city applications across multiple domains. Throughout we describe how the proposed smart city platform supports extensibility and smartness, and how it has the desired attributes of replicability, scalability and interoperability.