Okay, I've got two comments this time.
One - the answer key appears to be wrong again. The article referenced is kind enough to explain that both 1000^8 and 10^24 are equal to 1,000,000,000,000,000,000,000,000 bytes, and that both 1024^8 and 2^80 are equal to 1,208,925,819,614,629,174,706,176 bytes. I won't try to check that (I'm good at doing calculations in my head, but not that good). I can't imagina that anyone here will not agree that 1,208,925,819,614,629,174,706,176 bytes is more than 1,000,000,000,000,000,000,000,000 bytes - so clearly, only two out of the four answers have to be checked, not all four. I'm quite sure that Steve will correct this as soon as night is over in his part of the world.
The more fundamental comment is about the question itself. It lacks context. "Largest unit of information defined" - by whom? for what purpose? when? (for I'm sure that one day, we'll need names for 1,000 YB and 1,024 YiB). But the whom and for what purpose questions are what I really miss in this question. A quick google search for "largest unit of information" learned me that the largest unit of information is actually:
* A module or chapter (http://blogs.ittoolbox.com/eai/implementation/archives/documentation-design-document-blocking-16698) - in the content of documentation design
* A bit (http://web.njit.edu/~walsh/powers/bits.vs.bytes.html) - in the content of units of information a computer can handle
* A "word" (http://oak.cats.ohiou.edu/~piccard/mis300/pcintro.htm) - in the context of what can be transferred between memory and CPU in a single step
Going back to SQL Server, I didn't even need to use Gooogle to come up with
* 2 GB - in the context of what can be stored in a single column
* 8,000 bytes - in the context of what can be stored in a single page
* 32 KB - in the context of disk I/O (extent size)
* Collection - in the context of the XML data type
In short todays QotD was
* Poorly worded (as it fails to give enough context to the question),
* Not related to SQL Server at all
* Not related to database design and theory at all
As a fix, I'd suggest Steve to change the question to "How much data can a YB store (check all that apply, for any definition of YB that is in use, recommended or not)". With this wording, the question is clear (though still not quite relevant for SQL Server or for databases in general for the foreseeable future), and the current answer key becomes correct.
On a final note, as someone else already mentioned up-thread, the remark in the referenced article that "Yottabyte can, in theory, store everything in the entire universe", is of course plain nonsense. I learned at school that 1 gram of hydrogen contains approximately 6 * 10^23 atoms (Avogadro Constant - though Wikipedia just tells me that this has since been redefined as 6 grams (grammes?) of pure carbon). Anywway, if I want to store just 2 bytes of state information for each of those atoms, I'd already need 2 YB. And I think that the entire universe holds just a tad more than a single gram of hydrogen or 6 grams of carbon...
Hugo Kornelis, SQL Server MVP
Visit my SQL Server blog: http://sqlblog.com/blogs/hugo_kornelis