-----Original Message----- From: Starlink [mailto:starlink-bounces@lists.bufferbloat.net] On Behalf Of Mikael Abrahamsson Sent: Wednesday, June 30, 2021 11:33 AM To: Dave Taht Cc: starlink@lists.bufferbloat.net Subject: Re: [Starlink] 69,000 Users On Wed, 30 Jun 2021, Dave Taht wrote: > I don't think data caps are needed. fairness is needed. [RR] Engineers always think fairness is "needed". Corporate managers who make the real decisions couldn't care less about "fairness". It's all about income - expenditures which is simply PROFIT! When fair becomes profitable, you'll see it. Until then,dream on. LTE has pretty good airtime fairness (but FIFO per customer). Still doesn't work very well when there isn't enough airtime to satisfy demand. [RR] There will NEVER be enough supply (of information carrying capacity in the network) to meet demand all the time. The reason is simple, until the supply is exhausted, there's gold in them thar hills and the carriers are not going to leave it there. They will fill their pipes until they burst, knowing that there is little if anything their customers will do about it until their pain threshold is exceed, and they spend a helluva lot of money on psychologists to let them know where that breaking point is and they aim to get there ASAP. Why?? PROFIT and MARKET CAP! Business 101! > I would prefer a solution that just billed for usage over a minimum. Well, data caps is similar to this. People generally don't like to get billed automatically upon higher usage, thus data caps can be used and people will have to go to some self-service page and "pay more" if they're over the cap, to get a higher cap. [RR] The people in charge at the carriers hate caps . they want customers unknowingly pouring cash into their coffers. CAPs came about when consumers complained to the regulators, not one day before! CAPs are nothing more than a compromise that forces the carriers to let their customers know when they are about to get sc___d. All communication networks ultimately reach an "uncomfortable equilibrium" where all parties are unhappy, but not unhappy enough to quit the game. Thinking that there is some technical means by which 1Gbps can be shoved through a pipe whose maximum possible throughput is only a fraction of that data rate is misguided thinking. More importantly, fairness is in the "eye of the beholder". One man's fairness is another man's inequity. The point is that any algorithm for handling packets in a congested/overloaded network at best will satisfy those that run the networks and not disappoint their customers "too much". It's bloody obvious that finding the algorithm or algorithms that achieve these two goals is the only path to success. RR -- Mikael Abrahamsson email: swmike@swm.pp.se