As you may have noticed, I’ve been posting less research/analysis in November and December than during some other periods. In no particular order, reasons have included:
- Over a 20 week period, I had travel in 13 of them.
- 3 of those were vacation in November.
- As travel finally wound down:
- It was time to focus a bit on my own business
- Elder care got serious; e.g., my parents went to the hospital on consecutive days, Christmas week, the first one on their 52nd wedding anniversary
- Linda and I both got really nasty colds
- The holidays were happening
- I started helping out a really cool startup company (first time I’ve taken stock in a private company in years; more on that soon)
- There was less industry news going on anyway than in some other recent months
But of course I plan to speed up the research/analysis/writing soon. Here, FYI, are a few things I have on my plate.
For a couple of years now, the center of what I’ve written about has been high-performance analytic data processing. You can expect me to keep pursuing that in all its aspects. But there are two specific areas I’ve identified in which I want to redouble my efforts.
First, almost every BI vendor has an effort in “in-memory analytics” and/or “interactive data exploration.” I suspect there’s a lot of difference in underlying technologies, but I’m having trouble getting details. QlikTech (the worst foot-dragger of the three), Microstrategy, and Jaspersoft all owe me follow-up conversations with the people who know what’s going on well enough to explain it. Tableau keeps promising me a briefing and then not delivering. And I’m even further behind with the behemoth companies — Oracle, Microsoft, IBM/Cognos (arguably) et al.
Second, solid-state memory is coming to data warehousing. The obvious reasons are that it’s obviously close, and Moore’s Law still applies to bring it closer. More specific reasons for believing in solid-state include:
- Teradata has made large strides in making solid-state memory useful.
- The stealth start-up I mentioned above is poised to make further strides.
- (I’m not totally sure yet about this part) The in-memory analytics mentioned above might wind up working better in solid-state memory than in DRAM.
I’m spending quite a few cycles thinking about this area.
I’d also like to look further at analytic applications and advanced analytic functionality. I foreshadowed some of that in my Aster webinars. There’s some good stuff to talk about at Teradata I should try to write up soon. I need to have a follow-up conversation with fascinating anti-fraud guy I met at Netezza’s London event. But that’s all just scratching the surface.
Both the MySQL and PostgreSQL communities are in some disarray. Other non-behemoth OLTP/general-purpose DBMS seem to be, at best, thriving niche products. (I see little in the way of innovative new use for, say, Progress, Cache’, Ingres, or anything multivalue.) But it feels as if there’s more opportunity out there than is being met. And at a minimum, I’d like to learn more than the almost nothing I know about OLTP NoSQL alternatives.
I’ve already said that I expect to give an industry-overview talk at MIT on January 28. I also have an overviewy press article and overviewy white paper under discussion. If those come to fruition, I’ll of course let you know.
Besides the above, I of course have a number of specific posts that I need to get around to researching and writing at some point, often on topics I’ve already written about before. Three subjects fairly high on the priority list are scientific data management, machine-generated data, and Oracle Exadata.
And finally, I have some subjects queued up for a couple of my other blogs as well. If you don’t already take our multi-blog integrated feed, this might be a good time to switch over.