CSR: Small: Collaborative Research: Scalable Fine-Grained Cloud Monitoring for Empowering IoT

点击翻译按钮获取中文摘要

We are in the age of unprecedented connectivity, where everything from computers, buildings, vehicles, medical devices, to even animals is being equipped to interact with each other by exchanging data over the Internet. This massively connected world, referred to as the Internet of Things (IoT), will play a crucial role in enabling never-before-seen computing services such as self-driving cars and personalized health care, consequently improving the quality of life. A key enabler for IoT is the cloud computing model that is already transforming computing in academia, enterprise, and the IT industry by enabling easy and on-demand access to computing resources and services at reasonable cost. Most current and future IoT applications cannot function without the infrastructure provided by cloud computing serving as the brain for controlling and processing the data from the IoT network. Thus, there is an urgent need for ensuring that the data centers that support the cloud services can sustain the ever-growing IoT needs, while maintaining high quality of services and meeting service level agreements. To this end, this project aims to undertake the fundamental task of collecting and detailed fine-grained monitoring of the health and usage statistics of rapidly growing data centers, and addresses the challenges therein.Fine-grained data center monitoring entails that the cloud providers have to acquire, transport, store, and process exponentially increasing amount of management and monitoring data. With the amount of such data increasing at an unexampled rate of over 10 billion+ records per day, conventional acquisition and transportation schemes are no longer sufficient and efficient to handle the data tsunami. This project aims to address the problem of efficiently acquiring and transporting cloud management and monitoring data. For data it will focus on efficiently moving the data from the points it is collected inside the data center to the points where it needs to be stored for processing. Successful completion will result in tools that cloud providers can use to monitor their data centers and obtain measurements useful to predict and/or localize performance bottlenecks. Finally, the educational and outreach goals of this project include training the next-generation computer scientists with skills to manage and design innovative IoT applications and systems.