Link to home
Start Free TrialLog in
Avatar of al4629740
al4629740Flag for United States of America

asked on

RDS farm vs. regular RDS ??

I hope I am asking a question that makes sense here.  I am looking to setup an RDS setup for my application.  What is the difference between an RDS farm and a regular RDS Windows server setup?  When would I need to consider one over the other?
Avatar of DevAdmin
DevAdmin
Flag of Italy image

In a RDS Farm you have multiple RDS Session Host for Hight Availability
Avatar of al4629740

ASKER

Lets say I would never have higher than 30 simultaneous connections for my application.  Would an RDS farm be something I should consider or is for more connections than that?

Also please define High availability.
ASKER CERTIFIED SOLUTION
Avatar of Brian Murphy
Brian Murphy
Flag of United States of America image

Link to home
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
Start Free Trial
It depends on the application.

I've had applications that were so CPU/Memory intensive that on server class hardware it supported 5 sessions each and yet the vendor still recommended Citrix/RDS because of the physical hardware requirement.

You would need to "load test" the application manually or automatically to ascertain the total # of users per RDS Server and even then highly dependent on physical hardware you are using and whether you are visualizing RDS as well.
High availability is redundancy.  If one hosting server goes down for maintenance/patching or unplanned their are others to continue making that application available.  There are different levels of HA.  If you have one physical server hosting multiple virtual machines then your HA is limited to the VM's.  This is a much longer conversation.
So would an RDS farm also have more security compared to a stand alone RDS?
I think there is an element called RDWeb.  Does that provide security on the farm configuration or does that come with RDS standalone also.
Many services can be deployed either standalone or as a 'farm'.

The term 'farm' is derived from the view of how a farmer sees his cows: replaceable.
In this deployment you generally create multiple servers that all perform the same function. The term 'farm' generally applies if they are deployed in a largely simple and repeatable way which makes each server very easy to replace and therefore largely unimportant.

This deployment usually also accommodates resiliency & load-balancing, but the 'farm' term traditionally means that any given server is effectively like cattle and is individually replaceable without impact.

You can deploy RDS comfortably and in a secure manner on it's own, across multiple servers or in a  'sever farm'. the constraints really relate to load. Have you performed any basic testing of what load your application places on any given server? if you can get a view of this it will help you work out how many servers you may need to handle 30 users. Additionally, have you got any requirements on DR/resilience in the event of a server failure?
With a farm you need sql server or equivalent to keep the users session variables that is configured on the connection brokders. This is done at the RDS Broker level. A farm also implies a Load Balancer for gateway/web servers. You also would want clustered storage


https://www.tech-coffee.net/deploy-a-windows-server-2016-rds-farm-in-microsoft-azure/
How would I find the load balance to distinguish between the two and what type of environment I should be configuring?
Two what?
What type of environment is entirely dependant upon what you are doing and on what hardware.
Do you need High Availability or is is just a nice to have feature?
Are you using session based, Personal Virtual Desktops, or Personal Virtual Machine? Going from left to right you increase the customization but also incur higher hardware costs as again from left to right the resource usage costs escalate.

For a single application it depends upon the resource usage of the application.. i.e. notepad and Adobe Premier use quite different resources. What is acceptable lag timing and what is unacceptable.

You could run RDS with one application on 2 servers (1 being the DC and the other holding all of the RDS roles needed  i.e. connection broker, Session Host, license manager. You could add RDWeb (support for multiple OS's) and/or RD Gateway (external access) if needed.