AnsweredAssumed Answered

FMS 16 Performance: 4 x remote sites

Question asked by edwardlscott on Jan 23, 2018

My client has 4 x sites around the country.  Their use of FM has evolved over the past 8 years from 2 x sites using FM Pro hosting 2 x local files that track memberships independently of one another, to (soon) 4 x sites.  The 4th site is new and is preparing to open soon.

 

With the addition of the fourth site, the client would like to now move to a remotely hosted FMS 16 configuration and access it from the 4 x sites.  The files could be allowed to remain stand-alone, one devoted to each site, or they could be combined into one file (by adding new site-sensitive infrastructure and importing all the data - this decision is not made).

 

The FM solution now in use involves the production of a 1 x page PDF that is stored to container fields (new members only).  The PDF creation involves communication between the local Mac hosts and a partnered-iPad in a choreographed (scripted) back-and-forth that enables a new member to review an agreement and to then sign it on the iPad, after which it is PDF'ed and stored to the Mac's FM file.  All this works nicely.

 

Point-of-sale performance is critical in a membership-type business as waiting in line for any reason is not good, and the local hosting of the files provides entirely adequate performance under the current architecture.  The client's targeted problem is that sites can not cross-check for member status when a member is out of town visiting another site.  Converting to a remotely hosted situation will unavoidable slow things down due to internet latency, the question being whether the slow down will be small enough to be tolerable.  I ran connectivity speed tests at the four sites, and here are the results summarized:

SitesPing (ms)Down (Mb/s)Up (Mb/s)
Site A1052.410.81
Site B3028.852.27
Site C43
60.4311.93
Site D1342.1611.47

 

In general I like these numbers, but the ping times are not great in places.  The upload speed at Site B is less than mind-blowing.  How significant are ping times in their affect on overall performance?  Based on the numbers above, and given that there are a lot of variables in the situation, what predictions can be made with reasonable reliability about performance (if any)?  How would you approach making the decisions for this project?

 

Thanks in advance,

Ed

Outcomes