Not only has most every article I've read on the subject of implementing a highly available Windows 2012 RD connection broker role has DNS round robin being used, but even where to enter the DNS name used to point to all of the connection brokers, is labeled"DNS round robin name", when going through the RD Connection Broker "configure high availability" configuration wizard.
Is DNS round robin a hard requirement, or can one use a hardware load balancer, which is an option for our organization?
Where I’m confused, especially when desiring a true highly-available connection broker role, is if a connection broker went offline, DNS round robin will still point new connection requests to it, so I’m assuming when an end-user “lands” (as DNS round robins through your connection brokers IP’s) on that connection broker, they’ll get a failed attempt, whereas as a load balancer could automatically take the IP out of the virtual IP name after x-number of failed health checks.
Thanks in advance for anyone who can clarify this, as I’m very confused as to why DNS round robin appears to be (at least from what I've read) the preferred method for "fronting" the connection brokers.