MS SQL Server 2008 database scaling

Question

I am trying to develop the best way to scale my site, and I have a question about how mssql will scale.

What the table looks like:

cache_id - int - identifier
cache_name - nvchar 256 - Used to search with event_id
cache_event_id - int - Basically, the grouping method
cache_creation_date - datetime
cache_data - varbinary (MAX) - data size will be from 2k to 5k

The stored data is a byte array, which is basically a cached instance (compressed) of a page on my site.

I see various viewing methods:
1) 1 large table, it will contain tens of millions of records and will easily become several gigabytes in size.
2) Several tables containing the data above, which means that each table will be 200 thousand. Up to a million records.

The data will be used from this table to display web pages, so anything more than 200 ms to get the record is bad in my eyes (I know that some ppl think that loading for 1-2 seconds is ok, but I think that it's slow and I want to do my best to keep it lower).

, , SQL-?
( )
?


, . , , , .

Answer 1

So it boils down to, what is it that slows down the SQL server?
Is it the size of the table ( disk space )
Is the the number of rows
At what point does it stop becoming cost effective to use multiple 
       database servers?

- " "; (, , ) 2- , , , .

, , , -, .. , , - , , , . / , .

. , ( - , , , , , , ),

( , ), ( ). , , DBA .

Answer 2

.

, , , .

1 . , , .

Answer 3

- , , . , , , .

, , . . ( , . .)

And consider the fact that your database design will evolve. This is not as scary as your comment suggests that you think so. It is much easier to change the database than the software that bypasses it.