Teradata Virtual Storage
One of the big features of Teradata 13.0, announced this week (Edit: and to be shipped some time in 2009), is Teradata Virtual Storage, which sounds pretty cool. So far as I can tell, Teradata Virtual Storage has two major aspects, namely:
- Teradata has engineered its optimizer to take serious accounts of differences between kinds of storage. This fits the “multitemperature” data warehouse story. (One caveat: I’m not totally clear as to which parts of this are new within Teradata 13.0, and which have been in the product longer.) Important differences in storage type include:
- Small, expensive, fast-spinning disk vs. big, cheap, slow-spinning disk. (I’m not too clear on the basic differences between KINDS of disk, like SATA, but I do understand the difference between 7,200 RPM and 15,000 RPM.)
- Interior vs. exterior of disk. (Same RPMs and 2X the effective circumference = 2X the data transfer rate.)
- Disk vs. solid-state (this is still a future use).
- Teradata boxes now automatically redistribute data according to usage patterns. Data that is used more frequently goes to faster storage. They key word there is “automatic.” You can’t do anything to tell the system that data with dates in the past 30 days will get the most usage, with second-most usage for data that’s 365-395 days old. Rather, one or several days after data is added to the Teradata system, Teradata notices that it’s being used a lot, and moves it to some place that gets particularly fast access.
I’m squirming a little bit at that fully automatic aspect, for two reasons. First, there’s the delay in moving new data to the hot areas. Of course, a simple workaround for that problem is for data to start out in high-speed “hot” areas, and only get moved to cooler ones if it’s not much used.
My second reason for squirming is that all systems exhibit pathological behavior from time to time, and it’s comforting to be able to override that, and I don’t know what kinds of manual override Teradata offers. I’ll see if I can get some answers before leaving Las Vegas.
Comments
3 Responses to “Teradata Virtual Storage”
Leave a Reply
[…] Teradata announced Release 13.0 of its software, with lots of new features. I’ll write about some of those in separate posts as well. (E.g., Teradata Geospatial, Teradata Virtual Storage) […]
The solid state disk (SSD) is the future — Nano-tubes or quantum or whatever — and languages that were “born parallel” to take advantage of the hardware. This is why “SQL RULES!!” to quote some graffiti ..
[…] ascribes considerable importance to a Virtual Storage technology whose main purpose is to allow mixing of heterogeneous storage devices in a single […]