An Overview of the Workshop
Let's start with the goals of the workshop that Mark laid out at the beginning of the class.
- Provide an overview of the available models for deploying FileMaker solutions across a WAN.
- Understand how different aspects of your database design can impact performance.
- Discuss how changes in the underlying server and network architecture can impact performance.
- Learn the importance and value of testing.
One of the comments from Mark during the workshop sums up the importance of this class, even if you think you may know a good portion of the information that will be presented.
Originally Posted by Mark Richman, owner and president of Skeleton Key
So the workshop is not focused on which technique is the fastest (takes the smallest unit of time), but instead on used bandwidth. The concept is simple...a technique will always perform faster if it needs to pull down less data from the server. Instead of debating whether a type of Loop is faster than some other technique, you know that with whatever your chosen technique, it will finish quicker if it only needs to download 200kb of data as opposed to 3MB of data.
Mark thanked Jon Thatcher (Lead Engineer on the Database Server Team for FileMaker, Inc.), for his information and expertise that lead Mark to these discoveries.
How Does the Workshop Flow
After introductions, Mark does a brief overview of the information that he presented at DevCon 2010. He goes over some of the tests that they have run. Tests that show how Portals (sorting / filtering), data modeling, and startup routine affect data transfer between FileMaker Server and client.
Then you get some simple examples of how to reduce the amount of data FileMaker needs to pull from the server. And I'll admit, while I knew most of this, there were a few of things that made me say, "What, really???".
It is at this point that the workshop becomes even more awesome. Mark takes a solution that each of the 4 participants provides, and shows you in your own file what things are causing excessive data download...and how you could fix it. And the solutions that he showed during the workshop on April 26 were super simple. Simple to the point of taking less than 5 minutes to diagnose and tweak to get better performance. Team that with some basic changes to your development habits, and you can see some impressive results coming from that file that you were ready to trash or completely overhaul. One example that Mark showed us, from one of the files provided, reduced the load time of a busy layout and intensive script from 30+ seconds down to about 11 seconds. Need I say more?!
Another killer advantage, Mark shows you the tools to measure and test for yourself. No more guessing about how much data is moving during any particular action. Real data, really awesome.
Impressed? Definitely!
Typically, I am not one to be easily impressed...but I am super-excited about what I learned in that class. It's time to move from 'database theory' to 'understanding your database clearly'.
Skeleton Key has this workshop scheduled a few more times over the next several months. If you develop solutions that run over the WAN, you should seriously consider this workshop. If you don't, well, most of the principles that are covered apply equally to working over a LAN and even a local file on your own machine.
For a full list of the classes offered by Skeleton Key, visit their website: SkeletonKey.com
