Embedded Databases for the Internet of Things

物联网连接了数十亿台设备，这些设备收集、存储和分析有关我们的环境、商业和工业流程、健康和活动以及我们的社会的数据。 开发和部署这些设备是一个价值数十亿美元的行业，所使用的技术及其功能多种多样。通常，这些设备的硬件资源非常有限，为它们开发软件是高度定制的，耗时且容易出错，因为软件是专门为每个硬件平台和应用用例开发的。其目标是研究技术，以显着降低成本和时间来开发嵌入式设备的数据收集软件，并实施和部署这些技术作为开源软件供全球社区使用。第二个目标是在工业和环境监测应用中部署和评估数据收集系统，并与这些数据分析至关重要的领域的研究人员合作。科学的方法是调查和评估专为低资源、低功耗嵌入式设备设计的数据管理技术，这些技术可以针对闪存的硬件限制和独特功能优化其行为。我们以前的工作已经开发了LittleD，最小的，用于嵌入式设备的SQL关系数据库，运行在低至4 KB的内存，IonDB，用于嵌入式设备的键值存储，以及新颖的优化文件系统和NOR闪存的闪存技术。这些技术和贡献将被完全集成和扩展，为嵌入式设备生成最小、最节能的数据库，内存在1到100 KB之间，代码空间不到256 KB。目前在这个空间中没有最小的数据库SQLite的系统，对于这些设备来说太大了。我们的目标是让IonDB成为一个无处不在的开源系统，用于像SQLite这样的嵌入式设备，用于手机。这项研究的重要性在于，它解决了一个迫切的工业需求，即数据库在当前没有系统可用的嵌入式设备上工作。该系统将使嵌入式设备的开发和部署更加高效。我们的研究小组将继续与基洛纳市和葡萄园等合作伙伴合作，将该系统应用于灌溉管理和环境可持续性，以及环境，健康和工业数据收集的其他应用。

The Internet of Things connects billions of devices that collect, store, and analyze data about our environment, business and industrial processes, health and activities, and our society. Developing and deploying these devices is a multi-billion dollar industry, and there is a wide diversity of the technologies used and their capabilities. Typically, these devices have very limited hardware resources, and developing software for them is highly customized, time-consuming, and error-prone as software is developed specifically for each hardware platform and application use case.******The objective is to research techniques to dramatically reduce the cost and time to develop data collection software for embedded devices and implement and deploy these techniques as open source software for the worldwide community to use. A secondary objective is to deploy and evaluate the data collection systems in industrial and environmental monitoring applications and collaborate with researchers in these domains where the data analysis is critical.******The scientific approach is to investigate and evaluate data management techniques that are specifically designed for low-resource, low-power embedded devices that optimize their behavior for hardware limitations and unique features of flash memory. Our previous work has developed LittleD, the smallest, SQL relational database for embedded devices, that runs on as little as 4 KB of memory, IonDB, a key-value store for embedded devices, and novel optimized file systems and overwriting techniques for NOR flash. These technologies and contributions will be fully integrated and expanded to produce the smallest, most energy-efficient database for embedded devices with between 1 and 100 KB of memory and less than 256 KB for code space. There are no systems currently available in this space with the smallest database, SQLite, being too large for these devices. Our goal is for IonDB to be the ubiquitous, open source system for embedded devices like SQLite is for cell phones.******The significance of the research is that it addresses a pressing industrial need for databases to work on embedded devices where there is no current system available. The system will make it more efficient to develop and deploy embedded devices. Our research group will continue to work with partners such as the City of Kelowna and vineyards on applications of the system for irrigation management and environmental sustainability as well as other applications in environmental, health, and industrial data collection.