[Bloat] [iccrg] Musings on the future of Internet Congestion Control

[Sebastian Moeller]

Hi Michael,


> On Jul 10, 2022, at 22:01, Michael Welzl <michawe at ifi.uio.no> wrote:
> 
> Hi !
> 
> 
>> On Jul 10, 2022, at 7:27 PM, Sebastian Moeller <moeller0 at gmx.de> wrote:
>> 
>> Hi Michael,
>> 
>> so I reread your paper and stewed a bit on it.
> 
> Many thanks for doing that! :)
> 
> 
>> I believe that I do not buy some of your premises.
> 
> you say so, but I don’t really see much disagreement here. Let’s see:
> 
> 
>> e.g. you write:
>> 
>> "We will now examine two factors that make the the present situation particularly worrisome. First, the way the infrastructure has been evolving gives TCP an increasingly large operational space in which it does not see any feedback at all. Second, most TCP connections are extremely short. As a result, it is quite rare for a TCP connection to even see a single congestion notification during its lifetime."
>> 
>> And seem to see a problem that flows might be able to finish their data transfer business while still in slow start. I see the same data, but see no problem. Unless we have an oracle that tells each sender (over a shared bottleneck) exactly how much to send at any given time point, different control loops will interact on those intermediary nodes.
> 
> You really say that you don’t see the solution. The problem is that capacities are underutilized, which means that flows take longer (sometimes, much longer!) to finish than they theoretically could, if we had a better solution.

	[SM] No IMHO the underutilization is the direct consequence of requiring a gradual filling of the "pipes" to probe he available capacity. I see no way how this could be done differently with the traffic sources/sinks being uncoordinated entities at the edge, and I see no way of coordinating all end points and handle all paths. In other words, we can fine tune a parameters to tweak the probing a bit, make it more or less aggressive/fast, but the fact that we need to probe capacity somehow means underutilization can not be avoided unless we find a way of coordinating all of the sinks and sources. But being sufficiently dumb, all I can come up with is an all-knowing oracle or faster than light communication, and neither strikes me to be realistic ;)


> 
> 
>> I might be limited in my depth of thought here, but having each flow probing for capacity seems exactly the right approach... and doubling CWND or rate every RTT is pretty aggressive already (making slow start shorter by reaching capacity faster within the slow-start framework requires either to start with a higher initial value (what increasing IW tries to achieve?) or use a larger increase factor than 2 per RTT). I consider increased IW a milder approach than the alternative. And once one accepts that a gradual rate increasing is the way forward it falls out logically that some flows will finish before they reach steady state capacity especially if that flows available capacity is large. So what exactly is the problem with short flows not reaching capacity and what alternative exists that does not lead to carnage if more-aggressive start-up phases drive the bottleneck load into emergency drop territory?
> 
> There are various ways to do this; one is to cache information and re-use it, assuming that - at least sometimes - new flows will see the same path again.

	[SM] And equally important, that a flow's capacity share along a path did not change be other flows appearing on the same path. This is a case of speculation which depending on link and path type will work out well more or less often, the question then becomes is the improvement on successful speculation worth the cost of unsuccessful speculation (mostly the case where the estimate is wildly above the path capacity). Personally I think that having each flow start searching achievable capacity from the "bottom" seems more robust and reliable. I would agree though that better managing the typical overshoot of slow start is a worthy goal (one if tackled successfully might allow a faster capacity search approach).

> Another is to let parallel flows share information.

	[SM] Sounds sweet, but since not even two back-to-back packets send over the internet from A to B are guaranteed to take exactly the same path, confirming that flows actually share a sufficiently similar path seems tricky. Also stipulating two flows actually share a common path say over a capacity limiting node we have say 99 other flows and our parallel already established flow in equilibrium, now our new flow probably could start with an CWND close to the established flows. But if the bottleneck is fully occupied with our parallel established flow the limit would be 50% of the existing flow's rate, but only if that flow actually has enough time to give way...

Could you elaborate how that could work, please?

> Yet another is to just be blindly more aggressive.

	[SM] Sure, works if the "cost" of admitting too much data is acceptable, alas from an end users perspective I know I have flows where I do not care much if the overcommit and start throttling themselves (think background bulk transfer) but where I would get unhappy if their over aggression would interfere with other more important to me flows (that is a part why i am a happy flow queueing user, FQ helps a lot in isolating the fall-out from overly aggressive flows mainly to themselves).


> Yet another, chirping.

	[SM] I would love for that to work, but I have seen no convincing data yet demonstrating that over the existing internet, however we know already from other papers, that inter-packet delay is a somewhat unreliable estimator for capacity, so using that, even in clever ways requires some accumulation and smoothing, so I wonder how much faster/better this is actually going compared to existing slow start with a sufficiently high starting IW? In a meta criticism way, I am somewhat surprised how little splash paced chirping seems to be making given how positive its inventors presented it, might be the typical inertia of the field or an indication that PC might not yet be ready for show-time. However, if you think of somethnig else than paced chirping here, could you share a reference, please?


> 
> 
>> And as an aside, a PEP (performance enhancing proxy) that does not enhance performance is useless at best and likely harmful (rather a PDP, performance degrading proxy).
> 
> You’ve made it sound worse by changing the term, for whatever that’s worth. If they never help, why has anyone ever called them PEPs in the first place?

	[SM] I would guess because "marketing" was unhappy with "engineering" emphasizing the side-effects/potential problems and focussed in the best-case scenario? ;)

> Why do people buy these boxes?

	[SM] Because e.g. for GEO links, latency is in a range where default unadulterated TCP will likely choke on itself, and when faced with requiring customers to change/tune TCPs or having "PEP" fudge it, ease of use of fudging won the day. That is a generous explanation (as this fudging is beneficial to both the operator and most end-users), I can come up with less charitable theories if you want ;) .

>> The network so far has been doing reasonably well with putting more protocol smarts at the ends than in the parts in between.
> 
> Truth is, PEPs are used a lot: at cellular edges, at satellite links… because the network is *not* always doing reasonably well without them.

	[SM] Fair enough, I accept that there are use cases for those, but again, only if the actually enhance the "experience" will users be happy to accept them. The goals of the operators and the paying customers are not always aligned here, a PEP might be advantageous more to the operator than the end-user (theoretically also the other direction, but since operators pay for PEPs they are unlikely to deploy those) think mandatory image recompression or forced video quality downscaling.... (and sure these are not as clear as I pitched them, if after an emergency a PEP allows most/all users in a cell to still send somewhat degraded images that is better than the network choking itself with a few high quality images, assuming images from the emergency are somewhat useful).

>> I have witnessed the arguments in the "L4S wars" about how little processing one can ask the more central network nodes perform, e.g. flow queueing which would solve a lot of the issues (e.g. a hyper aggressive slow-start flow would mostly hurt itself if it overshoots its capacity) seems to be a complete no-go.
> 
> That’s to do with scalability, which depends on how close to the network’s edge one is.

	[SM] I have heard the alternative that it has to do with what operators of core-links request from their vendors and what features they are willing to pay for... but this is very anecdotal as I have little insight into big-iron vendors or core-link operators. 

>> I personally think what we should do is have the network supply more information to the end points to control their behavior better. E.g. if we would mandate a max_queue-fill-percentage field in a protocol header and have each node write max(current_value_of_the_field, queue-filling_percentage_of_the_current_node) in every packet, end points could estimate how close to congestion the path is (e.g. by looking at the rate of %queueing changes) and tailor their growth/shrinkage rates accordingly, both during slow-start and during congestion avoidance.
> 
> That could well be one way to go. Nice if we provoked you to think!

	[SM] You mostly made me realize what the recent increases in IW actually aim to accomplish ;) and that current slow start seems actually better than its reputation; it solves a hard problem surprisingly well. The max(pat_queue%) idea has been kicking around in my head ever since reading a paper about storing queue occupancy into packets to help CC along (sorry, do not recall the authors or the title right now) so that is not even my own original idea, but simply something I borrowed from smarter engineers simply because I found the data convincing and the theory sane. (Also because I grudgingly accept that latency increases measured over the internet are a tad too noisy to be easily useful* and too noisy for a meaningful controller based on the latency rate of change**)

>> But alas we seem to go the path of a relative dumb 1 bit signal giving us an under-defined queue filling state instead and to estimate relative queue filling dynamics from that we need many samples (so literally too little too late, or L3T2), but I digress.
> 
> Yeah you do :-)

	[SM] Less than you let on ;). If L4S gets ratified (increasingly likely, mostly for political*** reasons) it gets considerably harder to get yet another queue size related bits into the IP header...


Regards
	Sebastian

*) Participating in discussions about using active latency measurements to adapt traffic shapers for variable rate links which exposes quite a number of latency and throughput related issues, albeit for me on an amateurs level of understanding: https://github.com/lynxthecat/CAKE-autorate

**) I naively think that to make slow-start exist gracefully we need a quick and reliable measure for pre-congestion, and latency increases are so noisy that neither quick nor reliable can be achieved, let alone both at the same time.


***) Well aware that "political" is a "problematic" word in view of the IETF, but L4S certainly will not be ratified on its merits, because these have not (yet?) been conclusively demonstrated; not ruling out the merits can not be realized, just that currently there is not sufficient hard data to make a reasonable prediction.



> 
> Cheers,
> Michael