Highly Available TurboFloat Server

Hello,

We're using TurboFloat for our on-premises software. We're beginning to roll out a "high availability" feature for our software but are now realizing that we aren't clear on how to make the TurboFloat licensing server also be HA (to avoid having a single point of failure in the system).

Can you provide any guidance on what technical requirements exist (is a standard HTTP load balancer in front of multiple server sufficient?) or any "best-practices" you'd encourage for such a configuration?

Thanks,Jeff Allen

Hey Jeff,

Right now (as of the time I'm writing this -- version 4.0.x), the TF Library <-> TurboFloat Server only communicates over raw TCP (basically binary blobs over TCP).

In the next feature version (4.1.x) we've added the ability for the TF Library and TFS to talk over HTTPS (meaning you can put HTTPS servers and proxies in front of the TFS).

And shortly after 4.1.x is released we're also releasing a version of the TFS hosted on our servers that is high speed, high availability, and available to you and your customers with the click of a button (no need for them to install it on their hardware).

Thanks, Wyatt.

With raw TCP have you seen customers successfully deploy a load balancer in front of TFS? Any recommendations on software or configuration (e.g. buffer sizes)?

Is there an estimated timeline for the 4.1 release?

Unfortunately, many of our customers are in offline/disconnected scenarios so I'm afraid we'll need to continue hosting these TFSs on-prem for the foreseeable future.

Thanks,Jeff

There are a number of tools out there to do TCP load balancing. Nginx is one such tool: https://www.nginx.com/resources/admin-guide/tcp-load-balancing/