2 Offices, 1 Server

[Mr. Sparkle]

We are an office running Practiceworks on a local server. Currently we have a satellite office remoting in over a directly connected Fiber connection. Lag is substantial from the satellite office.

We are exploring best solutions to migrate our PMS, and are wondering whether there is a cloud version of OD. And if not, whether renting server space from, say, AWS and installing OD on there would be a viable option.

We have digital radiography and digital pans, no CBCT.

Any advice would be greatly appreciated.

Thank you,
Mr. Sparkle

[agrasley]

Hello,

There are multiple ways that you could handle this with Open Dental. I will list them below with a few quick pros and cons. I would recommend contacting support to go through these options in more detail if you are interested!

* Open Dental Cloud
- we do offer a cloud solution, which could be a good option. This does require a constant internet connection to be functional.

* Cloud server (AWS, RackSpace, etc)
- can scale well, and be fairly performant. Requires typically higher ongoing cost, added complexity with local devices such as imaging hardware and printers/scanners.

* Local servers at each location with database replication.
- Can be very fast, as each location has a local server, does not require constant internet (although spotty internet can be problematic). Adds complexity to setup, could require more support time when/if things go down, some chance of data loss when things do go down.

[Ray.Casey]

I am new to this and have been thinking through some physical/logical desings. From what I can tell the biggest bandwidth requirement will be for file shares and not client/server DB queries. We use XDR and OpenDentImages. I plan to create file cache servers that are local to the offices (like \\localserver\opendentimages and \\localserver\xdr) and host the mysql on a OD VM running in the our Azure cloud instance that is part of our M365 cloud tenant. Then I will use something like OneDrive or Azure file sync to mirror files shares. Essentially its a file cache very similar to how cache servers work on the internet. It may only apply yo our eccosystem.

[Ray.Casey]

Why would the DB's TCP transactions and payload require data base replication or two phased commits? And how do ya queue transaction logs for long running processes or CRUD operations? Am I incorrect in assuming that the heavy payloads and latency issues are file shares and file i/o vs db transactions?