carloscastilla - Fotolia
A lot has changed since SQL Server 2005 launched 11 years ago, and DBAs who plan to upgrade following the end of extended support on April 12 are about to get a taste of what's new.
Bala Narasimhan, vice president of products at PernixData, and David Klee, founder and chief architect of Heraflux Technologies, believe that SQL Server 2005 end of life could be a boon to DBAs. SearchSQLServer discussed with Narasimhan and Klee the trends to watch as database development advances faster than ever.
This is the second part of a two-part article. Click here for part one.
What is your advice for DBAs upgrading after SQL Server 2005 end of life?
David Klee: I like to go to the latest major version. It's the licensing level of it that depends. I say, sit back and look at the business requirements around it -- availability, performance, features. Look at where the application and the business are going with it. Where's my scalability going to be three years down the road? Is the volume doubling every year or is it staying fairly flat? Look for what you [will] need then. Look for what you need now. And try to weigh price versus features.
David Kleefounder and chief architect of Heraflux Technologies
Bala Narasimhan: In addition to that, I would also consider technology evolution. The last decade has seen changes. You've seen storage evolve from spinning [disks] to flash. You've seen servers go from 65 GB of RAM to, today, easily being able to support more than a terabyte of RAM. You've seen CPUs with 20 cores become standard. There are the same changes on the networking side. The cloud is another interesting trend. I would [consider] those technology trends as well and combine them together.
What trends should DBAs pay more attention to as they change their SQL Server implementations following SQL Server 2005 end of life?
Narasimhan: It seems to me that the application world -- which includes the relational database, SQL Server, Oracle, all of that -- [has] been in this silo. People who write applications and people who manage databases have been doing that one way, and the infrastructure folks -- which include the storage admin and the virtualization admin -- they are doing their thing as well. ... If you could bring them together, there are a lot of benefits that the data center can exploit. Getting the infrastructure folks to appreciate applications and what they do more, and for the applications folks to understand what goes into the infrastructure that supports them is definitely very interesting.
Klee: A lot of people talk performance, performance, performance. ... I'm starting to shift my view away from performance directly and onto efficiency. ... I want to make this more efficient at all levels of the application stack and infrastructure; by doing that, you get better performance out of it. But, we've hit a time in the industry where too many people have been throwing hardware at the problem for too long without getting to the root cause of it.
One good thing about the end of life of SQL Server 2005 is it forces people to revisit some of these architectural decisions. If we can make things more efficient as part of the upgrade process, it can actually cut down on the demands on the infrastructure below it, and everything gets faster. Business scales and everything gets better.
Narasimhan: One interesting development over the last few years, on the data side at least, is the fact that big data technologies have become mainstays. In my mind at least, the best use case for those big data technologies is on the data center itself.
Imagine a world where you could connect data on all of your applications and all of your infrastructure and run data analytics on all of that. Then you could optimize the data center for the application and the application for the infrastructure.
Today's analytics are very siloed. There is analytics above the infrastructure and analytics above the application. Getting those two to work together and building analytics that connect the two will be very interesting because it leads to what Dave was talking about, where you can make the infrastructure work for the application. Sometimes you overprovision the application or you break them because you don't know what the capabilities of the underlying infrastructure are. So now you can tune the applications as well, so it's a two way street. I think that applications will help drive that.
How to handle an upgrade after SQL Server 2015 end of life
Everything you need to know about SQL Server 2014
Microsoft's Analytics Platform draws on both SQL Server and Hadoop