This is an interesting conversation, and I wish I had more time to contribute to it atm.
A couple of questions that I have not been able to test/document that perhaps others would know.
<>what is the average/peak bandwidth consumption of a AC client machine while in active use? (e.g. opening a chart, updating an encounter, viewing an imported item)
<>what is the average/peak bandwidth consumption of a AC client machine while in active use running remotely? [e.g. LMI, remote desktop, terminal)
<>has anyone tested AC running in a cluster environment to determine stability/data integrity in that architecture? [lic costs aside]
The four primary factors to balance are performance, data concurrency, reliability, and price.
In answer to question David, one of the sites that we support as a manged installation has 4 practitioners, 22 client machines, ~ 7000 patients, and one tired server.
We have continued to groom the network, tweak the existing mish-mash of hardware, and phase in new Win7Pro boxes as funds allow. The office manager is proficient in restarting machines, logging into the server from clients, and they only require our assistance about once a month, and system-wide outages are now rare.
Most commonly, if we have to get involved, we do some quick checking, but generally just restart the server if it hasn't been otherwise restarted [because of maintenance, MS patches] in a couple of weeks. Piss-poor from my enterprise perspective, but 'fastest/bestest' way to resolve. For contrast, we also run Unix/Linux servers for other clients that have two years + of up-time, and we hate to restart them, but power maintenance/outages are usually the culprit. Otherwise the servers would be still up from 2005 when we installed them. Part of the MS environment/constraints.
And yes Ketan, I know you want me to add a sig, and I'll get A-Round-Toit. :P
Indy
[not a Dr]
[Technologist Emeritus]
[of a company that I don't plug here]
