By submitting your email address, you agree to receive emails regarding relevant topic offers from TechTarget and its partners. You can withdraw your consent at any time. Contact TechTarget at 275 Grove Street, Newton, MA.
McGehee started working with SQL Server 11 years ago. He is founder of the community site SQL-Server-Performance.com, as well as author of Professional SQL Server 2005 Administration, among other technical books. McGehee is also currently the director of DBA education for Red Gate Software and is a frequent speaker at industry events such as SQL PASS and SQL Connections.
SearchWinIT.com: If an IT shop has just migrated to SQL Server 2005, or is still on an older version, is there any compelling reason to upgrade to version 2008?
Brad McGehee: Microsoft has essentially rewritten Analysis Services and Reporting Services from scratch and made [the product] much more scalable. With Reporting Services, for example, now you don't need IIS. That's a big thing because IIS is kind of a pain in the neck to deal with and it's a security risk. It also meant more maintenance and management and more things breaking down. By getting rid of IIS and having everything done natively, things can be a lot easier to manage, are more efficient and more secure.
My general advice, though, is if your [SQL Servers] are working well and performing well, [then] don't upgrade. Upgrading is an expensive risk, so why rock the boat.
How do you think SQL Server 2008 makes life easier for IT managers and administrators?
McGehee: The Management Framework in 2008 lets you control SQL Server's new policies. In [SQL Server] 2005 version or earlier, you can set up a server any way you want. If you have 53 different servers, you can have 53 different configurations, which is a real pain in the neck from a maintenance or management perspective.
Another new management feature [in SQL Server 2008] is Performance Studio. You can use Performance Studio to manage and track the performance of one server at time. You don't have a central control panel to see the performance of all your servers though. For IT managers, [Performance Studio] may be interesting, but it's not a compelling reason to move to 2008 because Performance Studio isn't enterprise-wide, whereas the Management Framework component is.
For an IT manager, Performance Studio still involves individual setup, maintenance, backup, performance tuning -- that means spending time moving things over, converting and setting up the backup schedule and maintenance plans again.
PowerShell is now part of SQL Server 2008. Do you think IT will find it a useful tool?
McGehee: I think it will be mostly ignored. Some DBAs are developers at heart, but they are in a DBA position. Those kinds of people might like it because they like to code it rather than do it by GUI. But for the average DBA who's not a developer, it won't be interesting. It's still too hard. Sitting down and writing scripts is very time consuming.
So you don't think a lot of DBAs are ready for PowerShell?
McGehee: I go to a lot of conferences and user group meetings, and one thing I discovered is that a lot of DBAs are novice DBAs. SQL Server is growing so much that DBAs are [becoming] hard to find, so they're being pulled from anywhere. Many DBAs know just enough to keep things working.
Is Microsoft doing anything to simplify the product because so many SQL DBAs are novices?
McGehee: Microsoft may disagree with me, but the product is getting more difficult and more complex all the time because they keep adding more features. [The features] are all interrelated quite a bit and the user interface is not for the faint of heart in [version] 2005. And in [version] 2008, it's even more complex with more features and options. By putting out the CTP now, [Microsoft has] given people [time] to start playing with it. Even now that's still tough because people haven't moved to 2005 yet.
Where are your clients in their SQL Server adoption?
McGehee: Based on my personal experience when I talk to people, I would say [fewer] than half of them have upgraded to 2005 yet. I don't know how representative that is of the real world, but most of them have not upgraded and the ones that do go to 2005 [do so] because it's a new install of SQL.
Is SQL Server on par with IBM and Oracle in the enterprise?
McGehee: In terms of reliability and performance, they can do just as well as IBM DB2 or Oracle can do. There's obviously some differences in platforms and features, but if [an IT shop] is not limited by hardware, based on capabilities and licensing costs, Microsoft has everyone else beat.
Based on following SQL since version 7.0, how is the product evolving?
McGehee: More enterprise-wide management. Microsoft has a lot of other products, add-ons for enterprise management, for example. I think the problem is that it's not going to be very successful because you need to build enterprise capabilities into the product, not have [them] as an add-on. The new Management Framework is a step in that direction and PowerShell is in that direction too. There are a lot of third-party products that make a lot of money off of managing SQL Server because it's a one-instance, self-contained product that doesn't play well with others. It's not easy to manage enterprise-wide. Eventually [Microsoft] will build more of those features in, but it seems to be a slow process.
Where does that leave the third-party SQL Server vendors?
McGehee: What Microsoft has included is a lot better than what it used to be, but it's nowhere near what third parties already provide. Backup compression is built into the enterprise version of SQL Server 2008, but Microsoft didn't include encryption. There are third-party products that do both backup compression and encryption and it's enterprise-wide rather than instance-based.
Do you think Microsoft is on track with its release date for the product?
McGehee: It will be awfully tight and they're behind on their initial announcements. So, based on Microsoft's own schedule, they're behind. I don't think the average DBA really cares. They'd rather see it come out late and be perfect versus early and buggy and have to wait for the first service pack before [they] want to use it.