Hey team in another thread @burke was helping me to understand how constrained some deployments are with regard to network connectivity. One of the effects of that is needing to support old configs that are difficult to upgrade. Which got me to thinking.
Could the implementation community benefit from a physical OpenMRS appliance? What if we e.g. configured an Intel NUC ($200) with two 1TB SSDs ($150 each), one for data, one for operations and updates that could be shipped back and forth.
Zero touch platform deploy anywhere in the world for $500, is there a use case?
There was some effort to do this using BRCK years ago. There was also a cool appliance designed for very rugged environments (maybe for MSF?) shared at a lightning talk at an OpenMRS conference years ago. I don’t know of current efforts in this space.
Just a base system, any peripherals would be more. Updates would be handled by periodic distribution/swapping of an new SSD, I realize there are some operational aspects I am not costing out either, like who configs the disks that are swapped.
In general I am trying to think how we can stop the number of configs we need to config/test/support from becoming unmanageable.
Following this thread. We are having this same discussion on software and support in limited areas. Some of the solutions mentioned are new to me, so will be interested to see how this evolves.
Connectivity is the largest challenge to many of the solutions we look at, mostly due to the operating environment including support, footprint size, and obstacles.
I’m following another thread on scanning of forms, into OpenMRS, which has even been another solution we’ve considered. We want to try and make our technology solution as like paper or a non-tech fallback solution as possible. Nothing crashes a mission more than have tech fail you and then switching to paper completely changes all the processes that the team is used to.