SQL Lite Speed

Introduction

Basic review. Product installed flawlessly and quickly. Less than a minute once I located the license key. Once installed, there is no program to run. Instead, there are a few simple extended stored procedures that you use to take advantage of this product.

First step, benchmark a backup of my databases for full backups

I am currently using a maintenance plan, so I have some statistics, but I decided to run Profiler and Performance Monitor and gather more information. The first test was to perform a backup of 25MB db. Here are the results:

Method	Options	Backup File size	Avg CPU (Min-Max)	Backup Throughput	Time
SQL Server Native	Default options	16787kb	5 (0 - 18%)	1,719,247 bytes/sec (max 13,763,625)	30sec
SQL Litespeed	defaults	1796kb (holy criminy)	10, from 1.5 - 43%	851,929 bytes/sec (max 17,890,513)	18
SQL Server Native	Default options	993MB	5 (0 - 15%)	3,416,286 bytes/sec (max 6,623,440)	2:35
SQL Litespeed	defaults	126MB (Amazing)	12, from 4 - 40%	6,048,520 bytes/sec (max 13,378,083)	1:02

After the backup is complete, the stored procedure reports data like so:

SQLLiteSpeed Version 2.6.04 (Release).
Registered Name: SQL Server Central 
Licensed processors: 0
Encryption Support: Yes
Processed 51928 pages for database 'epo', file 'ePO' on file 1.
Processed 1274 pages for database 'epo', file 'ePO_log' on file 1.
BACKUP DATABASE successfully processed 53202 pages in 28.412 seconds (15.339 MB/sec).

One thing I noticed with further testing, in QA, the query for the backup took 1:12 to run, though the time is shown as 28 seconds. Not sure of the problem. This same operation, according to my maintenance plan logs, took 1:38, so this database went from 405MB to 105MB and was quicker with Litespeed! The vendor has acknowledged some discrepancies with the times reported. I suspect they are capturing only the time to grab the data pages and write them out, rather than the entire stored procedure time.

How does this sound? Pretty good to me, so I took the next step after a few days of letting this run on my test system.

The Real Deal

Now it's time to see how this really works. I moved this to a production system after a week of testing and found the following:

Method	Options	Backup File size	Avg CPU	Backup Throughput	Profiler CPU	Profiler Reads	Profiler Duration
SQL Server Native	Default options	447.70MB	12.8	3,416,286 bytes/sec (max 6,623,440)	500	724	127440
Litespeed	default	119.26MB	22.5	6,048,520 bytes/sec (max 13,378,083)	374	389	36610
Difference to SQL		-328.44MB	+10%	126ms less	435 less reads (??)	90.83ms less
Litespeed	Compression=2	105.9MB	17%	5,864,970 bytes/sec (max 24,834,778)	471	389	40214

Truly amazing statistics. The backup savings alone are worth the price of this product.

Restores? Working great. I made full, differential, and log backups and was able to restore each of these without an issue. I even tried restoring out of order and Litespeed caught all my tricks, stopping each restore and forcing me to restore them in order.

In addition, this product works with the same options as the native SQL system. Recovery, Norecover, etc. are required when restoring multiple logs. The syntax is slightly different, but the options are the same. And the procedure is the same, full backups, last differential, logs in order.

Still nervous? The last thing I tested is the "extractor" software. This allows you to convert a SQL Litespeed backup into a Microsoft Tape Format (MSTF) device. I ran the conversion (a command line software program) and it extracted my 2MB Litespeed backup into a 20MB file that I used to restore from Enterprise Manager. I should note that this was an encrypted Litespeed backup that was extracted to an unencrypted MSTF file. I had to supply the encryption key for the extraction, but the resulting MSTF file easily restored using Enterprise Manager with no password.

Conclusions

I am a huge proponent of using the native SQL tools to backup to disk and then your normal backup system to get these files on tape. I have used Arcserve, BackupExec, and other agents for SQL backups at times, but haven't trusted them completely, found them to be somewhat unreliable, and have preached the gospel of using native tools for backups.

Then along comes Litespeed and makes me rethink myself. This product works. Fantastically, and in a large environment (dozens of servers), this can pay for itself many times over in less disk space, less network traffic, and most importantly (because they are hard dollars), less tapes.

It isn't often I find a product that I like and would use, but this is just such a product. In fact, I am testing it further along with my own replacement for the maintenance plan that will manage my backups and files for me. I urge you to give this product a try and see for yourself.

Rating

Return on Investment	5.0 - Well worth the cost for any size company in resources.
Ease-of-Use	4.0 – It's easy to use, but the lack of any GUI or automatic scripting integration makes it slightly cumbersome when restoring lots of files.
Features	5.0 - What more can you ask. It does everything the native BACKUP command does, only better.
Learning Curve	4.5 – Got to know how to script, but that's it.
Documentation/Tutorial	4.5 – the only thing you need are some more sample scripts for managing the backups.
Time savings	5.0 - Way faster than the native backup.
Lack of Bugs	5.0 - None found during this review.
Support	5.0 - Support provided through email for the review. The few questions I had were answered within hours, regardless of when I sent them.
Overall	5.0 – There’s nothing like this on the market!

Specifics

Vendor Information

DBAssociates Pty Ltd

Phone: +61 3 9640 0992

Address:

Level 2, Charter House

4 - 10 Bank Place

Melbourne VIC 3000, Australia

E-mail : info@dbassociates.com.au

Website : http://www.sqllitespeed.com

Pricing

SQL LiteSpeed

Price : Contact Imceda :

Volume discounts are also available. Contact dbassociates for more details.

An evaluation version is available at http://www.sqllitespeed.com/slsevaluation.asp

SQL LiteSpeed - Lightning Fast Backups