Notifications
Clear all
Topic starter
I was wondering what everyone else is doing with their infrastructure. I have a SAN and Blade array setup but both are nearing their end of life as dictated by the vendors, unfortunately. A year or so ago, there was an edict at a conference of "Don't buy any new hardware" and that word got back to me from the people who attended the conference. Even if CPS was in the cloud by now I still need hardware for other things to run this business.
So, would anyone like to share if they have a SAN/Blade setup? I currently have a EMC VNX 5400 SAN and a Dell M1000e Blade array. I love them both and they are still sufficient for our business but the vendors will eventually stop supporting them and I am already paying a premium to keep this "old stuff". If this stuff does eventually get into the cloud I won't need a high end SAN and might be able to go with something lower end.
I know some other clinics I have talked to use some Dell SANs, and several are using Cybernetics. My vendor would love for me to buy HP, probably because they have a bigger margin on the sale but I don't want performance problems or problems in general. I don't have much experience with anything but EMC SANs or Dell Blades.
In a perfect world I would like to keep them going for more than 5 years without the usual upgrade coercion but I know that may not be possible.
Mike Zavolas
Tallahassee Neurological Clinic
Posted : September 5, 2019 8:24 am
Hi Mike,
There are a ton of options out there. As far as the HPE route Nimble is now a part of that portfolio and I would highly recommend a Nimble array(s) for a SAN. I have several in production without any issues for many years now. We pair the Nimbles with Dell PowerEdge servers but that is also due to the fact we have used Nimble for many years prior to HPE purchasing them.
For the blade center replacement of course you could just go to another blade center, but I would look into something more like 1u servers. Gives you more flexibility down the road to replace the servers or add/remove them as your needs change.
However there is always going away from a SAN and Blade setup and going to something like a Dell EMC VXRail setup using converged infrastructure. This has Compute, storage and networking all wrapped into a tightly integrated setup leveraging VMWare and Dell EMCs management and automation engines.
Needless to say the options are pretty endless today and to give you a more sound recommendation I would need to know more about your infrastructure, workloads and sizing requirements.
-Rob
Posted : September 5, 2019 5:36 pm
We're running HP servers and SAN under VMWare and very happy with them. We will probably be replacing most of the infrastructure next year do to age. Looking at moving to the cloud but currently it would double our IS operating costs, could be different for you as we receive non-profit licensing from Microsoft being an FQHC.
Steve
Posted : September 9, 2019 1:38 am
Second Nimble. If you are a Dell shop - Compellent if you can afford it or equallogic as option B. Hyperconverged things like Nutanix may be of interest too - compute/storage in one box (no first hand experience with that).
Cloud seems like an inevitability and might as well include it in the decision and pricing...at first blush it will be more expensive...especially given the requirements of CPS on how much memory and vcpus you need to have depending on number of sessions...BUT, it is a service that you can scale back once CPS/Athenahealth is fully web based. Depending on time commitment (1-3 years) with AWS, Azure, etc, you will get substantial price reduction if you purchase reserved instances...
Posted : September 11, 2019 9:24 am
Topic starter
Thanks for the recommendation. I heard about Vx Rail recently from someone else and will look into that as well as other options.
Posted : September 13, 2019 5:23 am
Topic starter
I have been hesitant to jump into the cloud thus far and I figure if I wait long enough it will be cloud based anyway (SaaS). My main concerns are the fact that we are not running thin clients. We do have Citrix access to some hospital resources and people have the usual thin client issues like printing and not logging out correctly. I supported Citrix in the past and have experience with the pitfalls, especially CPS related things which I see discussed here in the forum. Secondly, we live in a part of the country which has hurricanes, seemingly on an annual basis now. When we lose our internet we can not do business. I also have another building which is connected and they are down too. Our data center has redundant generator power, redundant cooling and the utilities are included in our rent. Our uptime has already taken a hit as cloud services have been unavailable (Centricity RT, Cspeak, and our ability to view hospital images). I guess I should start another thread but how are your clinics handling potential outages? I have looked into redundancy with Internet but that basically doubles our cost. I suppose if I had a thin environment I wouldn't need as much bandwidth and could cut things down to essentials for emergencies. We have had one outage earlier this year which was the result of a fiber cut several months ago. Other than that we have been good with this ISP since we have been with them over 7 years now, including multiple hurricanes. The cloud outages have always been Azure/AWS and whatever Athena uses for the RT plug in.
Someday I think it will all be SaaS. I would prefer that to a thin desktop situation. But when it goes out we won't be conducting business during the outage I guess.
Mike Zavolas
Tallahassee Neurological Clinic
Posted : September 13, 2019 5:35 am
Topic starter
Thanks for the reply
Posted : September 13, 2019 6:26 am
I hear you.. We have a hub/spoke topology - Centricity is the hub in a datacenter with multiple redundancies and our sites are all over the state. Most have redundant (diverse paths) circuits back to the "hub". You are right that it increases the cost but in our case part of it is subsidized and operations have made the call that the cost of having secondary circuits is justified. Like you said, the second circuit doesn't need to be as beefy as the primary - just to get you by until service is restored. The main thing is to have them take diverse paths so one backhoe doesn't take out both :).
I have also seen and considered solutions that can get you buy in a case of CPS being inaccessible - A PC with software that downloads periodically and retains 4-5 days worth of patient information. That PCS + USB Printer are on emergency power. If your circuit gets cut you can revert to that and still see patients. Some business continuity at least...
As far as thin clients - I love the ThinOS ones from Dell - in exam rooms we have the all in one 5040 (there is a new version with larger display now) and we also have 5060 and 3040. We have Wyse Management Suite subscription and it makes managing these thin clients a breeze. Honestly it's the best investment (and it wasn't a big investment either) we've made. Usually order clients with ThinOS+PCoIP. This give you flexibility to change your broker to whatever you want - VMware, MS, Citrix...even aws workspaces if you decide to do so in the future.
Our back end is RDS - we publish apps (Centricity, IE, Citrix receiver for access to partner hospitals, etc.) in exam rooms and in other areas present full desktop.
Posted : September 13, 2019 8:16 am
