Devolutions Gateway

Hi,

We haven’t gotten to document that topology yet, although we use it ourselves.

We could have a quick session to go over your options, it’s quite simple, although we have a lot of improvements already planned for June.

Just contact us at service@devolutions.net when you’re ready.

Hi Maurice

Do you prefer by mail or here in the forum?

I was able to install the Devolutions Gateway via Devolutions Server Console on a standanlone Server :-)
Is it possible to install only the Devolutions Gateway without DVLS Console? Or do you reccomend to install in any case DVLS Console?

What sizing recommendations are there for the Gateway server?
How many sessions are supported at the same time?
We are 45+ people - how many Gateway are recommended?
What is the best way to loadbalance / failover the Gateway? (e.g. loadbalancer in front of the Gateway, how to configure)?

I'm building a new Environment (currently only with Trial-License, atm we only have a RDM Site-License and not yet DVLS-Licenses).
Do you recomment the latest stabel version (RDM: 2021.2.29.0 / DVLS: 2021.2.14.0 / Gateway: 2021.1.6.0 ?) or should I use the Beta?

Thank you!
Regards Felix

Hello,

• The console should be viewed as a tool not only for the DVLS Instance(s), but for the whole platform. We are making efforts to break inter-dependencies and have a lightweight console that is used for every distinct component of the platform : Instances, scheduler, gateway, and recording server.
• The Gateway server is really a impressive achievement for our engineering staff, 75 concurrent sessions is in our spec sheet, but we have a customer that goes 100+. The tunneling itself requires little resources, the most intensive part of the process is for TLS encryption for inbound and outbound connections. This means CPU could be a hotspot, way more then memory. Obviously, one should also monitor the network interface.
• After taking into account CPU and network requirements, the only other factor is really the network isolation that is needed. At Devolutions, we have a tiered system for isolating the various pieces of our network infrastructure, this results in using one Gateway instance per tier. It really depend on your desired isolation for various destinations.
• For Load balancing, the focus is support for TCP connections. That's where we need to document more and determine a baseline for the typical workload that we can support. We do use Microsoft ARR internally in our lab, but I've challenged the support staff to get familiar with NGinx because it should support a heavier workload just because if its nature.
• As for the proper mix of versions, for DVLS beta, the next version has many improvements, it will probably be the GA release published tuesday. As for RDM, the beta is pretty stable even now. For a lab, I would use them without hesitation.

You should know that a independent GW server will require a paid license, we can hook you up with good trial keys for the whole stack since you're going through an intense POC.

Best regards,

Hi Maurice

Thank you for your fast answer!

At Devolutions, we have a tiered system for isolating the various pieces of our network infrastructure, this results in using one Gateway instance per tier.

The tiering is exactly one case for us for different gateways - alongside with multiple customer environments (we are a Cloud / Service Provider).

Thx & Regards
Felix