Commr. McDowell in the Washington Post
Commr. McDowell’s opening lines sounded like a real thriller was coming:
The Internet was in crisis. Its electronic “pipes” were clogged with new bandwidth-hogging software. Engineers faced a choice: Allow the Net to succumb to fatal gridlock or find a solution.
So what happened in the year 1987?
The loosely knit Internet engineering community rallied to improve an automated data “traffic cop” that prioritized applications and content needing “real time” delivery over those that would not suffer from delay. Their efforts unclogged the Internet and laid the foundation for what has become the greatest deregulatory success story of all time.
What event - what Internet Crisis of 1987 - is Commr. McDowell talking about? A couple of lists I’m on spent some time trying to figure this out.
Hat tip to Scott McCullough, who pointed us to this thesis. Page 27 says:
‘The area of Internet congestion control was baptised in 1986-1987 when the then ARPANET suffered ‘congestion collapse’ [57]. Congestion collapse had been predicted by Nagel [89] in 1984. Congestion collapse occurs when mounting levels of traffic result in high packet loss inside the network, such that few or no packets are actually delivered to their destination, yet each link is highly loaded.’
‘The initial response to ARPANET’s congestion collapse problem was to increase the capacity of the network. This helped temporarily, but the ARPANET continued to suffer congestion collapses until a strategy to control the load of packets entering the network was developed. In 1988 Van Jackson enhanced the famous Transport control protocol (TCP) [57] so that the transmission rate was responsive to the level of network congestion. TCP was made to reduce the rate of transmission of hosts when it sensed the network load was nearing congestion collapse. Since the introduction of this enhanced TCP, congestion collapse did not reoccur.‘
So the problem was that TCP kept opening the window to allow more packets to flow out, and was insensitive to actual conditions - like the Sorcerer’s Apprentice story, where the wish to have help with the wash (or whatever it was) had no limits. Answer: fix TCP so that when it gets the feedback that packets are being lost it slows down the flow of packets. TCP is indifferent to the nature of the packets that are flowing - video, voice, or email, it’s the same.
But then another writer chimed in and said, No, in fact in 1987 there was an effort to program the original internet backbone, NSFNet, to give priority to interactive (Telnet) packets over ftp requests. Hmm. Another writer said that there wasn’t a word about this prioritization in any NSFNET report.
So, what’s the true story, o readers? And was Commr. McDowell exaggerating?
Well, we know he was when he said the internet was a great deregulatory story. The whole thing was only possible because there was regulation — the carriers were forced not to discriminate against data flowing across their last-mile access points, and had to allow ISPs to connect to them, and the ISPs were expressly exempted from paying some special charges. Anyway, yes, lots of regulation.
But was there prioritization in NSFNet?