Every few years, the database research community prepares a report listing the most promising research directions. The previous one was called the Lowell report, and I was inspired by it. The latest one is called the Claremont report.
Some bits I found interesting:
- There is a call to exploit remote RAM and Flash as persistent media, rather than relying solely on magnetic disk. Indeed, Solid State Drives are an order of magnitude faster than our spinning disks and large pools of RAM are becoming affordable. External-memory algorithms are no longer a hot topic? (Yes, it is not that simple.)
- Web 2.0-style applications bring new database workloads. I did some work on merging Web 2.0 and OLAP and can testify that the Social Web is a good source of new database problems.
- They recommend integrating compression and query optimization. My work on compression in bitmap indexes that there are still open issues regarding compression in databases. Mostly, whereas information theory has taught us much about how to optimally compress, we have learned relatively little about how to use compression to save CPU cycles.