Storing and Managing Bits

Instructor Steven Swanson Email: swanson @ cs.ucsd.edu IM (not email): professorswanson@{AIM, Yahoo!, google talk, MS Messenger} Office: EBU3B 3212 Office Hours: By appointment UCSD homepage

Course Description

The world is overflowing with data, but existing technology for storing and moving it into a processor cannot keep up. As the onslaught continues, new technologies (like flash memory and other solid-state memories) have emerged. They offer tantalizing performance gains and they will promise to bridge the gap between big, slow, non-volatile storage and small, fast, volatile memories. As is often the case, this advance in hardware technology requires a re-evaluation of how applications manage, allocate, and use the technology.

This course is will cut a swath across storage technology from silicon, electron spin, and spinning disks through hardware and operating system management layers and into applications. The goal is to gain perspective on existing approaches to storing, managing, and accessing data in contemporary technologies and understand how those techniques apply to emerging technologies.

We will cover a range of topics and will focus on less conventional solutions storage problems. In particular we will discuss (at least):

• SRAMs, DRAMs, and memory hierarchies.
• Flash memories and emerging non-volatile solid-state memories
• Hard drive technology (how disks work, error correction, RAID)
• Memory allocators (malloc and company)
• Volatile storage abstractions
• Persistent storage abstractions (file systems, relational databases, object-oriented databases)
• Large-scale storage systems (distributed file systems)
• Programming models for big data sets

The class will consist of student presentations and discussion. We will divide the topics in the course among the students. For each topic, the student will research the topic, select a set of readings, and present a lecture. The lectures should focus on how the underlying technologies affected the current solutions to these problems and how the approaches would change if applied to new technologies. Since some of the topics are covered in othe grad classes, we will focus on less conventional solutions to problems that arise in each area.

As part of the presentation you will create a wiki page on the course wiki. You should list links to the resources you used in your investigation, so that others can go poke around if they are interested. The wiki is hosted here. You will need a google account. Please make sure all the links work. If there resources you can't link to for some reason, email me the files along with a topic word, and I will put them online. They will be accessible as http://www.cse.ucsd.edu/classes/fa08/cse249a/resources/<topic>/<filename>, and you can link to them there. Please, make sure there are no spaces in the filenames. The wiki page for my lecture on DRAMs and SRAMs has examples of all this.

The Wiki

Schedule

I will post the slides for most lectures. Since the slides contain material I am not allowed to distribute publically, they are password protected. I have posted the username and password to the web board.