Best Filtering Strategy

[fishtech]

Hi,

I need to configure filtered syncs to some (not all) remote users. This is to prevent some users from seeing commercially sensitive data available to other users.

I am hoping someone can advise the appropriate strategy for me between 'Custom SQL Qualifier' and  'Mirrorsync Customization Script'.

I'm using JDBC right now. I could cut over to XML, but I think JDBC is probably right for me. Can someone please confirm?

I'm syncing a solution withe the following characteristics:

• 3 files, 3.5Gb total

• 12 tables

• 18 remote users sync to Hub

• Hub has 20 local users

• Remote (mirrorsync) users authenticate  same local FileMaker 'sales' account to access the solution

• Local users authenticate to directory services (Open Directory).

 

Many thanks,

ft.

 

Edited August 23, 2017 by fishtech

[Junior360Works]

Hey Fishtech,

  It sounds like you've already looked through our documentation, but just incase, I'm going to post a couple links here as a reference (both to relevant parts of the same Advanced Documentation site). If you go with JDBC, you'll end up using custom SQL qualifiers, and for XML you'll be using the customization script.

Regarding JDBC v XML performance: http://docs.360works.com/index.php/MirrorSync_advanced_topics#Performance_questions

Regarding Customization options: http://docs.360works.com/index.php/MirrorSync_4_advanced_topics#Customizing_MirrorSync

  In general, we tend to recommend JDBC. However, depending on how complex your filtering gets, JDBC slows down at a faster rate as filtering data requires more and more queries. It's hard to tell exactly at what point one becomes faster than the other, but if it looks like filtering is occurring with several tables, XML may be a safer bet. If you have the time, I'd encourage you to try both.

Should you have any further questions, please let us know and feel free to email us direct at [email protected].

Good luck,

Junior Perez, 360Works Support Team

 

 

[fishtech]

Many thanks for your reply.

Is the speed of response of the custom SQL qualifiers 'test' function related to potential sync-speed impact when filtering with SQL?

Thanks,

ft.

[Junior360Works]

Hey FT,

  Exactly right. The 'test' speed should be the same as when those qualifiers are running in a sync. Sounds like a great way to test out configurations options. I'll keep that in mind when users are deliberating between using JDBC or XML in the future.

Thanks,

Junior Perez, Support Team Member

[fishtech]

Sounds good, thanks for the info. The test speeds seem pretty zippy to me in early testing but I need to do need to build queries across several tables.

ft