1

Is there an easy way that any of you guys use to determine this?

Ok, we have more than x gb of data, we should go for IQ! Ok, we have more than x reports going, we should go for IQ! Ok, we have a lot of data archived and need better timings retrieving it, we should go to IQ!

What are the factors that you use to make this decision?

  • It seems to me that you are putting a solution before the problem. What is the problem you are trying to solve? Define the requirement. Investigate solutions to your problem that meet your requirement. Eg. If you just want reports to run faster adding ram and some performance tuning might be faster and cheaper than implementing a whole new product/system. – Sir Swears-a-lot Aug 16 '16 at 9:54
1

Tough call. Look at it from a transactional RDBMS versus reporting-style OLAP perspective, at a high level. The IQ methodology is column-store, versus a traditional row-store for ASE. Column-stores are very fast when running aggregated queries over them, but that depends on how your reports are run, right?

I think you almost nailed it with this:

Ok, we have a lot of data archived and need better timings retrieving it, we should go to IQ!

So, do your analysis on volume of data you'll be querying all round, that is fairly static or slowly changing. If that's going to be queried a lot, look into IQ.

Don't be scared of a hybrid solution either. I have several SQL Server customers that have massive warehouses running with columnstore indexes, right in the same cabinet as their highly transactional RDBMS.

  • Thanks for you answer :) I'll go for a hybrid solution for sure! – aF. Aug 16 '16 at 13:53
1

IQ can be mind boggling fast, but it's a weird creature. It's a mix of the ASA database server (super small and low maintenance, formerly used for embedded applications) and the high-end data warehousing column store db server. Sort of like mixing a tricycle with a Ferrari.

It takes some specialized admin skills (especially if you're using "multiplex" cluster like installations), and the licenses are expensive. If you have tbytes of data, you're doing data mining queries, and your management doesn't flinch at a few hundred thousand for licenses and associated admin personnel cost, go for it (eg, if you're a big bank or something)

Also, keep in mind, that Sybase has been bought by SAP. SAP tends to focus of big huge software packages and they have their own big huge HANA in memory db. I have no idea how well they'll support IQ.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.