LWN: Comments on "The burstable CFS bandwidth controller" https://lwn.net/Articles/844976/ This is a special feed containing comments posted to the individual LWN article titled "The burstable CFS bandwidth controller". en-us Sat, 01 Nov 2025 14:23:34 +0000 Sat, 01 Nov 2025 14:23:34 +0000 https://www.rssboard.org/rss-specification lwn@lwn.net The burstable CFS bandwidth controller https://lwn.net/Articles/846859/ https://lwn.net/Articles/846859/ scientes <div class="FormattedComment"> With SeL4&#x27;s multicriticality, processes can also pass their scheduling context around, and without broken priority inheritance.<br> </div> Sat, 20 Feb 2021 05:16:41 +0000 The burstable CFS bandwidth controller https://lwn.net/Articles/845518/ https://lwn.net/Articles/845518/ flussence <div class="FormattedComment"> So there&#x27;s still the elephant in the room - how does it compare to MuQSS after these patches? In terms of latency/interactive performance that&#x27;s been consistently embarrassing CFS for a decade and a half now, and the only reaction seems to have been “ignore it and hope it goes away”.<br> </div> Tue, 09 Feb 2021 20:14:49 +0000 Have you looked at how network QoS does it? https://lwn.net/Articles/845457/ https://lwn.net/Articles/845457/ ras <div class="FormattedComment"> Linux network QoS solves a similar problem. QoS tackles how to share a network link equitably among many users.<br> <p> The CBQ and HTB are queuing disciplines attempting to solve the problem of having a single link between protocols needing guaranteed low latency but low traffic (eg, VoIP and interactive like ssh), tasks that need to a responsive link with uneven loads (eg, http), and that just need heaps of &quot;low grade&quot; bandwidth (eg, email). They use ad hoc techniques like the ones described here, and mostly work on a good day - but sometimes don&#x27;t. And they are computationally expensive.<br> <p> HFSC came later and solves the same problem. It has a rigorous mathematical analysis behind it, delivers perfect results, and is computationally inexpensive. The key turns out to be how you pose the problem. Doing that in a way that allows you to come up with a robust solution is non-obvious, or at least I found it non-obvious. Interesting, like the proposed solution here HFSC must also take into account what bandwidth was used and went unused in the past to determine what can be used the future.<br> <p> Unfortunately CPU scheduling and QoS are only similar, not identical. QoS has the luxury of the application breaking the work it presents to the QoS scheduled into bite sized pieces - packets. The QoS problem reduces to deciding which of these packets to send next, and when. In CPU scheduling you have a number of tasks lining up to use the CPU that will run for an unknown amount of time. The problem reduces to &quot;how long can I let this task run, before I interrupt it&quot;. Nonetheless, I suspect they have one thing in common: in order to do their jobs well, the mathematical model behind them must be perfect.<br> </div> Tue, 09 Feb 2021 05:25:54 +0000 The burstable CFS bandwidth controller https://lwn.net/Articles/845439/ https://lwn.net/Articles/845439/ jmclnx <div class="FormattedComment"> Interesting, almost like a Subscription Service some Businesses have, carry over unused time. KaaS :)<br> </div> Mon, 08 Feb 2021 18:26:09 +0000