>>>>>>The customer does have multiple locations. As far as whether it will be one SQL Server database or each site will have its own, not yet decided. I want to stay away from replication since I don't know how it is done and don't have the experience. We just had a conference call to just have a preliminary discussion. I would like it to be as simple as possible.
>>>>>
>>>>>Hmm... if fundamental things such as shared vs separate SQL Server databases have not yet been decided then IMO it's too early to discuss low-level details. Regarding bandwidth and latency, just try to minimize traffic over slow WAN (Internet) links. When discussing bandwidth in this kind of situation, remember that uplink speed is often more important than downlink, but often it's slower on asymmetric connections such as DSL or cable e.g. 60Mbps down/12Mbps up.
>>>>>
>>>>>If everyone has separate SQL databases, and you can use a loader approach as Thomas suggests, do you even need to have a shared server VM at all? EXE and metadata updates could be on a web or FTP site instead.
>>>>
>>>>First, thank you for some numbers.
>>>>
>>>>You see, each site will only need to see their one data. So if I combine all sites into one database I will still logically separate them (by site ID). But I need some time to work on this approach and they (the customer) may not give me the time.
>>>>So, my preference would be to have a separate VM for each site, with the exe and meta data, as well as a copy of the SQL Server.
>>>>Then, I will just need to devise an approach if the corporate wants to see some combine reports, pulling some summary data across all databases.
>>>
>>>In that case it sounds to me like the "simplest" solution would be to have a server with your app components on-site at each location. That server could be a VM or a physical computer.
>>>
>>>It's only if you are contemplating hosting your app off-site from the individual offices, that you need to worry about bandwidth, whether you should be looking into RDS etc.
>>
>>Thank you. I have a newbie question. Is it expensive (approx how much) for them to spin a VM server? I am just trying to figure what it will cost them to spin all VMs for all sites.
>
>If you're going to use cool slang, the term is "spin up" :)
>
>It depends. The lowest possible cost would be if they already have a large-ish virtualization platform in place at each site, with spare RAM, CPU and disk capacity (i.e. no more hardware needed) and they've invested in a DataCenter version of Windows Server so they can spin up new Server VMs without having to buy any more Windows Server software licenses.
>
>Large environments may already have suitable VM templates set up so it's very easy to create a new VM for you. If that's not the case there will be IT costs for manually installing and configuring Windows Server on the new VM.
>
>Worst case would be they have to buy some more hardware, and another Windows Server licence. If they're not already virtualized at a given site that might mean buying a complete new server computer, installing and configuring it.
>
>In any case you should have an idea of exactly what your server needs to run, and how many resources it will need. At a minimum you might start with 2 vCPUs, 8GB RAM and at least 100GB disk space for a Server 2012 R2 or Server 2016 VM running an instance of SQL Server with a small-ish database. To make sure SQL Server runs well in a VM, Google [sql server on a vm].

Al, thank you very much. This is very helpful. These sites are hospitals (and not small ones). So I am assuming that they already have a lot of "components" that you outlined above. So I will suggest to them to put into the budget a VM server for each. This will be simpler for me and not that expensive for them.
As far as the resources they will need. I usually specify for my app: min of 2 GB of RAM, 50 GB of disk space and SQL, 2008 or later. And typically they put much more than my minimum.
Again, thank you.