11 ms·

Why the internet only just works (2006)

198 points by llamport 10y ago

powera 10y ago
Article is from 2006.
- 56245623456 10y ago
  And? It's still a very good paper. Age doesn't change that.
  - powera 10y ago
    First of all, the internet has changed in the past 10 years. Second, it's best practice here to put the date in the title for old articles.
    - 56245623456 10y ago
      > First of all, the internet has changed in the past 10 years. Not much. We're still using TCP/IP, BGP and NAT. Web-standards may have evolved, but the internet is still the same, except for the average link being 10x faster now. > Second, it's best practice here to put the date in the title for old articles. I agree.
    - jamez1 10y ago
      If you think the internet is the world wide web as well, then sure, it's changed a lot. But the actual internet infrastructure is still the same, this paper discusses the protocols and architectural problems. I challenge you to name something that has changed. Even IPv6 was formalized 15 years ago.
      - tzakrajs 10y ago
        One of the WWW things that has changed, maybe not Internet technology specifically: if you are a business that moves very large volumes of packets, it doesn't matter that you have plenty of upstream bandwidth because the residential ISPs will refuse to upgrade their equipment on behalf of their customers. See: http://www.businessinsider.com/netflix-comcast-deal-explained-2014-2 http://www.businessinsider.com/netflix-comcast-deal-explaine...
      - craftkiller 10y ago
        Does TCP Fast Open count?
alkz 10y ago
except it doesnt
- john_gulliver 10y ago
  With more than 3 billion people using it, it arguably does
jstimpfle 10y ago
Thanks, for me this was mostly another illustration how to make a website that doesn't. Please give me at least a chance to get a plain pdf (or html) that I can read (and click into) without content bumping forth and back and my computer grinding to a halt.
- van_gaal 10y ago
  This seems a bit harsh. As far as I understand the purpose of this website is to make papers easier to understand. I think it accomplishes that. They are not even trying to show you ads. how are you helping?... Also, they might have restrictions about distribution of papers. Youtube also doesn't have a download link next to every video.
  - jstimpfle 10y ago
    It might sound harsh to you, but you'd feel differently if you sat in front of my computer. Each misclick costs me 1-2 seconds of layout recalculation (100% on 1 core). For reference, my computer is AMD X250 from ~2011, 4GB RAM (enough), and running Firefox on Debian 8. Using chromium it's a tiny little quicker, but still annoying. They could easily stop the content bumping and the darkening "please subscribe" overlays.
    - van_gaal 10y ago
      sure. talking to people on the internet only just works...
      - jstimpfle 10y ago
        This was probably a witty reference to the article. Can't appreciate it since I can't bring myself to suffer the website...
    - merb 10y ago
      well I have a mbp late 2013 and it costs me 1-2 seconds (16 gb ram and core i7 3rd gen...)
Animats 10y ago
As one of the early workers on network congestion, much of what he says is right. We really have no idea how to deal with congestion in the middle of the network. The best we can do is have more bandwidth in the middle than at the edges. Fortunately, the fiber optic and hardware router people have done so well at providing bandwidth that the backbone has mostly been able to keep ahead of the edges. We never dreamed of a connection with over 10,000 packets in flight. Cutting the congestion window in half on a packet loss and ramping it back up one packet at a time was something I came up with around 1984. That does need to be rethought, and it has been.
- schoen 10y ago
  Hopefully HN readers know that Animats is the inventor of https://en.wikipedia.org/wiki/Nagle%27s_algorithm https://en.wikipedia.org/wiki/Nagle%27s_algorithm
  - matt_wulfeck 10y ago
    Every time he posts we identify him as the author of Nagle's and someone that's done pioneering research in TCP/IP. It would be great if HN had some sort of annotation for distinguished commentators.
    - msie 10y ago
      We would have the same problem as the black bar. But this is a great(?) idea for a browser extension...
      - bbcbasic 10y ago
        But not everyone would have the extension.
        taneq 10y ago
        Isn't that precisely the point, though? If you want it, you can have it. If you don't, then you aren't bothered by it.
mhandley 10y ago
Interesting to see this paper making HN. I think it has stood the test of time reasonably well - it was written in about four hours one morning under rather unfortunate circumstances - I was stuck on a train that hit a vehicle at a crossing. I'm interested in what HN folks think has changed. What are today's urgent problems, and what are just annoying ones that will cost time and money to work around, but aren't critical? How do they differ from those of 10 years ago? Arguably, some I discussed back then have only grown in importance, with no solution in sight - for example, DDoS attacks. Whereas some, like address space exhaustion, look like being adequately addressed, as IPv6 is finally rolling out (only took 20 years!).
- quanticle 10y ago
  I'm wondering if your opinions on the difficulty of introducing new protocols have changed over the past decade. For example, Bittorrent has taken off, and while I used to have to manually configure ports on my NAT to allow Bittorrent to pass through, I no longer need to do so. Universal PnP seems to have taken care of that (at the cost of some security).
  - mhandley 10y ago
    STUN/TURN/ICE have enabled NAT hole-punching to become fairly reliable, but the complexity involved is horrible. Try reading RFC 5245 and its companions - this isn't a sane way to design an Internet. It does work though. New protocols are still an issue, but the problems differ depending on whether we're taking about the transport layer or the application layer. We've tried and largely failed to deploy alternatives to TCP and UDP. My effort, DCCP, for multimedia traffic is dead in the water due in part to NAT/Firewall traversal problems, and is now essentially tunneled over UDP. SCTP is successful in some limited circles, but you're very unlikely to see it end-to-end across the Internet for the same reasons. Even TCP is really hard to extend. If you're interested, we did some work on this a few years ago: http://www0.cs.ucl.ac.uk/staff/M.Handley/papers/extend-tcp.pdf http://www0.cs.ucl.ac.uk/staff/M.Handley/papers/extend-tcp.p... It's amazing what some middleboxes do to TCP. At the application layer, it's somewhat easier. You can deploy new protocols over UDP, as BitTorrent did and as Google has done with QUIC, and they'll work in most places, but you'll still need a fallback. The universal fallback these days seems to be to tunnel over TLS on TCP port 443, because that's rarely blocked. The fact that we need to do this is an admission of failure on the part of the Internet architecture. The fact that it just about works is the reason we've not needed to address the underlying problem, which is the tension between security and extensibility.
canada_dry 10y ago
The author (Mark Handley)'s other publications: http://www0.cs.ucl.ac.uk/staff/M.Handley/papers/ http://www0.cs.ucl.ac.uk/staff/M.Handley/papers/ Some very good reading.
peter303 10y ago
In the early web days (mid-1990s) Bob Metcalfe(invertor of Ethernet) humorously the internet would collapse any day from increasing traffic. Aint happened yet.
- walrus01 10y ago
  Despite recent best efforts to the contrary by level3...
hNewsLover99 10y ago
Wow! 42 comments on an article titled, "Why the internet ONLY JUST works" (emphasis added), and the word "security" occurs once, and the words "malware", "privacy", "malicious", "ransomware", "hack" and "cybercrime" don't show up at all. Is there another internet out there that I don't know about? If not maybe the title should be "Why the internet only just stinks"!
walrus01 10y ago
Looking at it from a network protocol level is one view. From another perspective the internet (whether ipv4 or v6) is a complicated set of peering relationships between ASes and business entities. It takes clueful and experienced people with "enable" on the core and agg routers who also understand the economics of transit, peering, IXes, transport, dark and WDM systems to build a proper ISP. The internet is only as fragile as you build it. I have seen too many ISPs running really important stuff through POPs where everything is 1+0.

198 points by llamport 10y ago

powera 10y ago

Article is from 2006.

56245623456 10y ago
And? It's still a very good paper. Age doesn't change that.
- powera 10y ago
  First of all, the internet has changed in the past 10 years. Second, it's best practice here to put the date in the title for old articles.
  - 56245623456 10y ago
    > First of all, the internet has changed in the past 10 years. Not much. We're still using TCP/IP, BGP and NAT. Web-standards may have evolved, but the internet is still the same, except for the average link being 10x faster now. > Second, it's best practice here to put the date in the title for old articles. I agree.
  - jamez1 10y ago
    If you think the internet is the world wide web as well, then sure, it's changed a lot. But the actual internet infrastructure is still the same, this paper discusses the protocols and architectural problems. I challenge you to name something that has changed. Even IPv6 was formalized 15 years ago.
    - tzakrajs 10y ago
      One of the WWW things that has changed, maybe not Internet technology specifically: if you are a business that moves very large volumes of packets, it doesn't matter that you have plenty of upstream bandwidth because the residential ISPs will refuse to upgrade their equipment on behalf of their customers. See: http://www.businessinsider.com/netflix-comcast-deal-explained-2014-2 http://www.businessinsider.com/netflix-comcast-deal-explaine...
    - craftkiller 10y ago
      Does TCP Fast Open count?

alkz 10y ago

except it doesnt

john_gulliver 10y ago
With more than 3 billion people using it, it arguably does

jstimpfle 10y ago

Thanks, for me this was mostly another illustration how to make a website that doesn't. Please give me at least a chance to get a plain pdf (or html) that I can read (and click into) without content bumping forth and back and my computer grinding to a halt.

van_gaal 10y ago
This seems a bit harsh. As far as I understand the purpose of this website is to make papers easier to understand. I think it accomplishes that. They are not even trying to show you ads. how are you helping?... Also, they might have restrictions about distribution of papers. Youtube also doesn't have a download link next to every video.
- jstimpfle 10y ago
  It might sound harsh to you, but you'd feel differently if you sat in front of my computer. Each misclick costs me 1-2 seconds of layout recalculation (100% on 1 core). For reference, my computer is AMD X250 from ~2011, 4GB RAM (enough), and running Firefox on Debian 8. Using chromium it's a tiny little quicker, but still annoying. They could easily stop the content bumping and the darkening "please subscribe" overlays.
  - van_gaal 10y ago
    sure. talking to people on the internet only just works...
    - jstimpfle 10y ago
      This was probably a witty reference to the article. Can't appreciate it since I can't bring myself to suffer the website...
  - merb 10y ago
    well I have a mbp late 2013 and it costs me 1-2 seconds (16 gb ram and core i7 3rd gen...)

Animats 10y ago

As one of the early workers on network congestion, much of what he says is right. We really have no idea how to deal with congestion in the middle of the network. The best we can do is have more bandwidth in the middle than at the edges. Fortunately, the fiber optic and hardware router people have done so well at providing bandwidth that the backbone has mostly been able to keep ahead of the edges. We never dreamed of a connection with over 10,000 packets in flight. Cutting the congestion window in half on a packet loss and ramping it back up one packet at a time was something I came up with around 1984. That does need to be rethought, and it has been.

schoen 10y ago
Hopefully HN readers know that Animats is the inventor of https://en.wikipedia.org/wiki/Nagle%27s_algorithm https://en.wikipedia.org/wiki/Nagle%27s_algorithm
- matt_wulfeck 10y ago
  Every time he posts we identify him as the author of Nagle's and someone that's done pioneering research in TCP/IP. It would be great if HN had some sort of annotation for distinguished commentators.
  - msie 10y ago
    We would have the same problem as the black bar. But this is a great(?) idea for a browser extension...
    - bbcbasic 10y ago
      But not everyone would have the extension.
      - taneq 10y ago
        Isn't that precisely the point, though? If you want it, you can have it. If you don't, then you aren't bothered by it.

mhandley 10y ago

Interesting to see this paper making HN. I think it has stood the test of time reasonably well - it was written in about four hours one morning under rather unfortunate circumstances - I was stuck on a train that hit a vehicle at a crossing. I'm interested in what HN folks think has changed. What are today's urgent problems, and what are just annoying ones that will cost time and money to work around, but aren't critical? How do they differ from those of 10 years ago? Arguably, some I discussed back then have only grown in importance, with no solution in sight - for example, DDoS attacks. Whereas some, like address space exhaustion, look like being adequately addressed, as IPv6 is finally rolling out (only took 20 years!).

quanticle 10y ago
I'm wondering if your opinions on the difficulty of introducing new protocols have changed over the past decade. For example, Bittorrent has taken off, and while I used to have to manually configure ports on my NAT to allow Bittorrent to pass through, I no longer need to do so. Universal PnP seems to have taken care of that (at the cost of some security).
- mhandley 10y ago
  STUN/TURN/ICE have enabled NAT hole-punching to become fairly reliable, but the complexity involved is horrible. Try reading RFC 5245 and its companions - this isn't a sane way to design an Internet. It does work though. New protocols are still an issue, but the problems differ depending on whether we're taking about the transport layer or the application layer. We've tried and largely failed to deploy alternatives to TCP and UDP. My effort, DCCP, for multimedia traffic is dead in the water due in part to NAT/Firewall traversal problems, and is now essentially tunneled over UDP. SCTP is successful in some limited circles, but you're very unlikely to see it end-to-end across the Internet for the same reasons. Even TCP is really hard to extend. If you're interested, we did some work on this a few years ago: http://www0.cs.ucl.ac.uk/staff/M.Handley/papers/extend-tcp.pdf http://www0.cs.ucl.ac.uk/staff/M.Handley/papers/extend-tcp.p... It's amazing what some middleboxes do to TCP. At the application layer, it's somewhat easier. You can deploy new protocols over UDP, as BitTorrent did and as Google has done with QUIC, and they'll work in most places, but you'll still need a fallback. The universal fallback these days seems to be to tunnel over TLS on TCP port 443, because that's rarely blocked. The fact that we need to do this is an admission of failure on the part of the Internet architecture. The fact that it just about works is the reason we've not needed to address the underlying problem, which is the tension between security and extensibility.

canada_dry 10y ago

The author (Mark Handley)'s other publications: http://www0.cs.ucl.ac.uk/staff/M.Handley/papers/ http://www0.cs.ucl.ac.uk/staff/M.Handley/papers/ Some very good reading.

peter303 10y ago

In the early web days (mid-1990s) Bob Metcalfe(invertor of Ethernet) humorously the internet would collapse any day from increasing traffic. Aint happened yet.

walrus01 10y ago
Despite recent best efforts to the contrary by level3...

hNewsLover99 10y ago

Wow! 42 comments on an article titled, "Why the internet ONLY JUST works" (emphasis added), and the word "security" occurs once, and the words "malware", "privacy", "malicious", "ransomware", "hack" and "cybercrime" don't show up at all. Is there another internet out there that I don't know about? If not maybe the title should be "Why the internet only just stinks"!

walrus01 10y ago

Looking at it from a network protocol level is one view. From another perspective the internet (whether ipv4 or v6) is a complicated set of peering relationships between ASes and business entities. It takes clueful and experienced people with "enable" on the core and agg routers who also understand the economics of transit, peering, IXes, transport, dark and WDM systems to build a proper ISP. The internet is only as fragile as you build it. I have seen too many ISPs running really important stuff through POPs where everything is 1+0.