Dynamic QoS Management for Speedify: Seeking Contributors for an Innovative Solution!

[ClintDavis]

Currently, QoS/SQM requires us to set a fixed throughput rate limit for the entire connection in order for its fair-sharing mechanism to be effective.

However, Speedify is designed to be adaptive; links can go offline, or Speedify might apply rate-limiting if the link reaches its maximum bandwidth. This leads to the available bandwidth varying significantly.

This made me ponder if there's a dynamic way to tackle this and I'm curious to know if anyone else has encountered this or has any ideas!

I've asked Speedify team and they don't have anything currently that can be used to suggest the QoS speed values, there are reported historic speed values however once applying QoS values it would form a cap and Speedify wouldn't go above those limits.

Objective:

To develop an intelligent mechanism that periodically calculates the QoS Up/Down values for the Speedify bond and dynamically adjusts them accordingly setting those values back into the SQM.

Targets:

Maximizing the throughput by setting the highest possible rate limit that adapts to changes in available bandwidth.
Addressing Speedify's switching between speed and streaming (aggregation) modes, it's per packet stream, so perhaps a best guess percentage reduction when this mode is in place?
Incorporating knowledge of the state of each adapter being monitored by Speedify.
Adjusting when Speedify applies the set rate limiting due to exceeding the monthly bandwidth on a link.

Actions:

Share your insights if you have encountered this problem or thought about it.
Propose algorithms or methodologies that might be applicable here.