Innovators in Performance and Reliability Technologies
| Account | View Cart View Cart |

"The Importance of Database Defragging"

Complete Transcript of
Marcus Galindo – Diskeeper Interview
on Let’s Talk Computers
Host Alan Ashendorf
December 08 2007

Alan:  If you own a business you know that you will have lots of information. You will have lots of data. And that means that it’s especially important that those large databases always be optimized to provide peak performance. To tell us why that is so important, our guest today is Marcus Galindo, Spokesperson for Diskeeper Corporation. Welcome back to Let’s Talk Computers, Marcus.

Marcus:  Well, thank you for having me.

Alan:  We’ve talked in the past about defragging - how important it is for either a home system or even on a server to maintain file integrity and to make sure that we can get the quickest access to file. But, when we’re talking about databases, there’s a whole different way of looking at it, isn’t there?

Marcus:  Well there is. Databases tend to become fragmented at a much faster rate. Databases can be some of the most intense processes that go on, when it comes to hard drive fragmentation, because there’s a lot of data being removed, deleted and therefore more prone to fragmenting.

Alan:  If you’re looking at a large video file, once that file has been written to your hard drive, it’s pretty well going to stay in the same place. But once you have defragged it and moved it to a nice little corner of its own world, you don’t have to worry about defragging it again. But databases are constantly being accessed and therefore, every time you open up a record and change a record, you have now made it so that it’s "discombobulated"!

Marcus:  Well that can be very true, especially when you’re talking about adding information, because the more you add the more space it’s taking up on your drive, and of course if that happens, depending upon space availability, you may run into the issue of fragmentation.

Alan:  Also, when you delete a record in a lot of databases, just because you said you deleted it, it does not necessarily mean that you did delete it.

Marcus:  Well, when you delete a file, particularly in a database, the space that it occupied is often still there after deletion. What can happen is, that as you are adding in new data (if that data is larger than the omitted space) a certain trunk may be left there, but then the rest of the file can be put somewhere else physically on the drive.

Alan:  To give you an example, let’s say that you are working on a spreadsheet and you start adding more and more rows and depending on where those rows are going to be stored, it’s going to be bringing the whole spreadsheet into memory before you can see it. And it you have to seven or eight places on the disk, it’s going to get slower and slower.

Marcus:  That’s very true. The more times different areas of the disk have to be accessed, the slower the processes are going to be. Those can be search queries, those can be a number of actions that are taken during database administration.

Alan:  Those are small files, but when you are going into a business environment, you have large databases. You have SQL databases that can be hundreds of thousands of records. When this gets fragmented, it really can slow down processing, can’t it?

Marcus:  That’s very true and those large SQL databases that you’re talking about are actually subject to several different types of fragmentation. There’s the "internal fragmentation", which has to do with a particular record being removed from a database page, in which case there’s that occupied space left over.

With internal fragmentation, one of the methods to use is the "DBCC DBREINDEX command". And that allows the rebuilt of clustered and non-clustered indexes. That’s going to solve the problem of internal fragmentation there.

Alan:  Okay, but what about "external fragmentation"? Where this database actually sits on your drive, [and] makes a lot of difference as far as how fast you can read records.

Marcus:  For external fragmentation, it really breaks down to kind of two different issues and that has to do with the actual file fragmentation, meaning the files that make up the database may be fragmented throughout the drive. And also there’s the "free space" fragmentation, which is the amount of unused space that could be used for adding onto the database.

Alan:  When we’re talking about a database, sometimes it is just one file. But in cases, it is a collection of files that work together as a unit. You can have the actual storage of the data and then you can have all the logging that goes with it.

Marcus:  A lot of times in the original databases, the log files are not fragmented. As the database grows over time, you will see that these files do tend to become fragmented, just based on the additional size they take up.

Alan:  I know you have a lot of feedback as to the difference a fragmented database, say a SQL database and one that has been properly maintained.

Marcus:  Internal and external defragmentation procedures can greatly decrease the amount of time it takes to run a lot of different operations on a SQL server. There are a lot of performance gains that can be seen through defragmentation. In these drives it’s particularly important, because you are accessing so many different files.

Alan:  Especially if you’re doing any kind of e-commerce, where a customer looking at a web browser is waiting for it to fill up. If it doesn’t fill up very quickly, what happens?

Marcus:  A lot of times, those customers can become frustrated with the site and simply go somewhere else.

Alan:  To a person who’s waiting for a screen to draw, seconds become minutes and they say, "I’m going to go elsewhere and go do my shopping." We’re in a world of "instant gratification", where you click a button and you expect something to happen immediately. And if your back-end is sitting there churning and churning and trying to get the information to throw it toward your front-end, the user is going to go away.

Marcus:  That’s very true. We’re in a day and age today where every second counts and making sure that customers get what they want as quickly as possible is certainly very important.

Alan:  How does Diskeeper help us to manage these large databases in order to help maximize our performance?

Marcus:  Diskeeper specifically works on the external fragmentation that’s seen in these databases. And what it does is it dynamically arranges your file fragmentation so that it’s defragmented in a way that makes sure that these files are easily accessed, quickly.

Further, it also does what we call, "free space consolidation" and "free space defragmentation". What that means is that as the database grows, because that free space is put into an optimized area, fragmentation is less likely to occur and at a much slower rate.

Alan:  What about "offline defragmentation"? We would pick a certain time in the day and say, "Okay, the system is being defragged." With e-commerce you’re looking at 24 hours a day, 7 days a week, every day of the year. And that doesn’t leave any time to be down to do any kind of maintenance. You have to be up all the time.

Marcus:  That’s why really "offline defragmentation" is no longer an option. And it’s no longer needed. With Diskeeper, what we have come out with is our new InvisiTasking Technology. What that allows us to do is defragment these databases on the fly, using only idle resources, so that these databases can stay up 24-7.

Alan:  What if somebody is using the database while we’re trying to defrag the database?

Marcus:  The number-one goal is to keep the system running as smoothly as possible at all times. That’s where the new InvisiTasking Technology comes in – is that it was designed specifically so that it can look at a computer’s usage, a server’s usage, and determine the best times to defragment. However, if during that defragmentation, some other process starts up, the defragmentation stops and allows you do what needs to be done and it will pick up during a more optimized time period.

Alan:  Will Diskeeper keep track where on the actual databases reside, so that it can maximize our throughput?

Marcus:  One of the new features included in Diskeeper is the IFAAST function. What that does is it allows frequently used files to be put into an optimized portion of the disk that is easiest to access. That means with this technology, if you have a database or particular files within that database that are used most frequently, they’re put into an area that can be accessed quickly. Therefore you’re going to have shorter "read times".

Alan:  How does a [Database] Admin work with the new Diskeeper in order to keep these databases defragmented?

Marcus:  A lot of Database Administrators can see, looking at SQL Server, for example, they may see “slow-downs”, they may see things like that. A lot of them are very familiar with familiar with internal fragmentation and using for example the "DBCC DBREINDEX" command to fix the internal fragmentation going on.

When it comes to the external fragmentation, Diskeeper actually takes over completely and is able to do everything in the background, on the fly, using only idle resources.

Alan:  A database is going to reside on the server and the two of them are kind of "married together", aren’t they?

Marcus:  Oh yes. A lot of times a particular server is set aside only for database administration. This is again where Diskeeper and defragmentation comes into play because if you have a particular server whose job it is to only manage the databases – file servers can become fragmented much faster than many other servers.

Alan:  I know that the new Diskeeper 2008 is basically "set it and forget it" and you don’t even have to even know that it’s there. But, Administrators like to see what’s under the hood and to take a peek to make sure that everything is running right now. From my viewpoint, it’s that that "warm, fuzzy feeling" is very important to me.

Marcus:  True and we have a lot people who share that viewpoint and because of that, we do have several reports that are available within Diskeeper that show different types of information. That is file performance, that can be the actual amount of file fragments, for those people who really want to see what’s going on.

Alan:  I finally realized that I really defrag my hard drive and I just got Diskeeper and I also do not have a lot of free space left of my hard drive to do defragging. How can the new Diskeeper 2008 help me in this scenario?

Marcus:  Well, if you’ve got a very low free space situation, our newest Diskeeper 2008 is actually specifically designed to handle issues of very high fragmentation and low free space. The engines that we’ve designed in this most recent version dynamically look at what is the best way to defragment your hard drive and goes about its business to defrag and keep your system optimized.

Alan:  Usually when you’re talking about utility software or maintenance software, you look at the cost of the software by itself. But when you’re looking at defragging and maintaining your system to run at its optimum performance, you’re not really looking at the cost of the software, you’re looking at the return of investment as to how much time you’re saving per each workstation, per each server, per each transaction, aren’t you?

Marcus:  That’s very true. You want to look at all the different angles in terms of the amount of money that you’ll save. And that is the amount of time that you’re going to spend manually defragging - the amount of time that you’re going to spend handling other issues. It can even be about the life of the hard drive, overall.

For a lot of corporations today, you’re talking about hundreds (sometimes even thousands) of hard drives, and when you add that up, and the savings that can be gotten over extending that life – even in only a year’s time period – you can be talking about hundreds of thousands of dollars.

Alan:  That is the bottom line as far as businesses are concerned. If somebody would like to find more information about defragging, where can they go?

Marcus:  They can go to our website www.diskeeper.com

Alan:  Marcus, it’s been our pleasure to have you as our guest here on Let’s Talk Computers, talking about how businesses can keep their computer systems running at peak performance by using Diskeeper 2008, with automatic defragging. We look forward to having you back on the air again real soon.

Marcus:  Thanks again for having me, Alan.