[dba-SQLServer] Rapid transactions

Let's imagine that there are 899 transactions per minute. Said transactions
infect/affect only one table of interest, but the data-flow is constant.
Let's also imagine that the boss wants to see numbers reflecting the state
up to the second.

Let's imagine that I build an OLAP db with everyting up to last year.
What I would want to do is somehow join the old data with the new data. I
already know how to do this but I'm looking for expert advice to point out
what could go wrong with my scenario. Let us suppose that I roll up the old
data about 5 times per day, and that a query as to where we stand now equals
"grab the old (consolidated) numbers plus teh current numbers. How might I
achieve this? Is it as simple as I think (roll up the old numbers and do a
join with the recent stuff and we're there), or am I overlooking something
important? I am assuing that anything that happened in 2006 and 2007 is not
subject to edit, and therefore can be rolled up in the classical AS format.
If I am correct, then T1 is Read0nly and T2 is the current stuff and a Union
get us to the current answer. Am I off-base on this conjecture? So far, I
don't think so, but I open the question for smarter minds than mine.

In the roughest strokes, we have an OLAP db and and OLTP. We have to unite
them with a single query. What has occurred a year or two ago is dead, so to
speak. It cannot be edited but it can be summarized using Group By etc. What
occurs this year is another set of data, and we have no choice but to do a
GroupBy etc. We want now to marry the datasets. IOW, if I want to computer
sales this year/month versus the last year/month, I need to union the OLTP
table with the OLAP table and then I have my result set. Is this correct or
am I missing something important?

TIA,
Arthur