Syllabus:
-
Background
This year, CS227 will focus on the data management needs of the Internet of Things (IoT). IoT is, loosely speaking, a term that covers any application that involves a large number of computers talking to each other and taking action without any human intervention. Examples include autonomous vehicles, automated manufacturing, location tracking and robotics. IoT will need to scale to many millions of devices, putting stress on conventional database systems.
If we look at the type of data that is most prevalent in IoT, we see that it is largely time-series. That is to say, it is a collection of values ordered by time. Update in place is rare; appending new values is common. We believe that in order to effectively scale this kind of data, new approaches are necessary. This is largely what this seminar is about.
We will read papers on topics related to IoT data and discuss them in class. Students will form teams and present the papers of the day to the rest of us. All students must read the papers in advance of class so that they can effectively participate in the discussion. Each student who is not a presenter should write a one or two page summary of the readings.
Students will also do a project. Again we will form small groups. Each group will work on a single project that is to be determined by the group and me. Projects could be:
- An application
- A component for a time-series data manager.
- A set of benchmarks
- Your best idea here
It is perfectly fine if two or more groups work on similar projects.
We will spend part of each class discussing these projects to give each team important feedback.