4 ms·

This never had any potential to decide the fate of the open internet. IETF generally documents what things are, not what some actors hope things to be. Countri

by Canada 4y ago

This never had any potential to decide the fate of the open internet. IETF generally documents what things are, not what some actors hope things to be. Countries that want sovereignty (control) over the internet can limit connection and participation to the extent they can, but they can't dictate what other countries that fully participate in the internet do regardless of what positions they hold at ITU or other organizations.

boltzmann-brain 4y ago
That's not really true. The ITU dictates what all sorts of telco equipment has to adhere to, and a bad actor could make it so that only compromised hardware (eg Russian or Chinese produced, with clear state involvement) can follow these standards. The IETF deals with a bunch of stuff, but that stuff runs _on top of_ what the ITU deals with. See here for more: https://news.ycombinator.com/item?id=33032452 https://news.ycombinator.com/item?id=33032452
- tialaramex 4y ago
  The ITU and IETF are very different. Although these are often lumped together as "Standards organisations" we should squint even at the word "Organisation" when it comes to the IETF. The IETF isn't an organisation - like the ITU or a Boy Scouts troop - it's an Activity, like dancing. Humans can participate in the IETF if they want, corporations cannot (corporations also can't dance) and since it isn't an organisation the IETF has no members and writes documents reflecting the consensus of participants. In this particular case an even more important difference is their approach to standardisation itself, the ITU produces "de jure" standards. The organisation produces documents which say how something shall be and that has the force of law. ITU members are sovereign entities, so they have police forces and armies, in principle they can choose to "enforce" the standards they've agreed, and only in practice do we end up relying on the fact that you might as well implement say I.762 as make up some other way to do it. The IETF produces documents how participants agreed things ought to work, but sometimes that's how they really work ("de facto" standards) and sometimes it isn't. For example the IETF published a document about how early browsers wanted Cookies to work, which briefly mentions how Netscape Navigator's first implementation did it, assuming this will quickly be superseded. It wasn't, the IETF RFC about cookies is only really useful today as either an aspiration or because it obliquely documents what you actually need to do, mimic Netscape (and guess which shortcuts work and which do not). The ISO and ITU approach works great where the main objective is to agree something and what it is exactly is secondary. For example the A-series paper standards from ISO are very nice, but if ISO had standardised what is now "US Letter" instead that would still be pretty useful even though it's an ugly shape without the nice properties of the A-series. However as you get into technical details this approach works much less and less well. In contrast the IETF approach works better for these details, because of the objective of consensus rather than an insistence on democracy, it will usually identify when there actually just isn't any agreement and standardise only what's actually agreed - the cookies example I mentioned is an example where this didn't work, but it's not that common. This has led to some conflict with "traditional" standards organisations which feel like they ought to just be able to write a document saying "No, it works like this" and have that take effect. Whether its ECMA, or the ITU, or even ISO this doesn't tend to go well for them but you can imagine it's frustrating even as they operate a liaison mechanism, to find that somehow the IETF is good at this while they can't comprehend how it even works.
- wmf 4y ago
  I would say the Internet has not run "on top of" telco networks for 10-20 years. Smart ISPs are running IP over MPLS over Ethernet over simple DWDM (OTN and even ROADMs are being dropped) which are mostly IETF and IEEE standards. When it comes to cellular, 3GPP appears to be upstream of ITU so we need to be vigilant against 3GPP getting corrupted but I haven't heard about that happening yet.
  - fweimer 4y ago
    It depends. DSL and GPON are ITU-T standards. DOCSIS and Ethernet are not. 3GPP (mobile networks) does not seem to be ITU-T-dominated either, although they reuse some ITU processes. ASN.1 and X.509 are widely used in some Internet protocols and are ITU-T recommendations. I don't know what WAN looks like these days, but there seems to be some ITU standardization in the WDM area. So it's rather mixed. I'm not sure to what degree ITU-T standardization of obnoxious features matters in practice. Not everything gets implemented. A lot of dual-use features are present on ISP networks in Western democracies because they are required by law. Institutional safeguards generally keep users from government harm in these countries, and not lack of standardization of the more invasive surveillance technologies.
    - throw0101a 4y ago
      > It depends. DSL and GPON are ITU-T standards. The IEEE has PON stuff: * https://en.wikipedia.org/wiki/Ethernet_in_the_first_mile#Passive_optical_network https://en.wikipedia.org/wiki/Ethernet_in_the_first_mile#Pas... Though most PON deployments seem to be ITU-based… for historical/inertial/first-mover reasons?… but that can be changed by throwing money at the 'problem' (should it become one).
- cryptonector 4y ago
  Yeah, but the ITU-T gets no say on what happens at the IP layer and up (and also at some layers below IP, like MPLS). They want a say, and they've been trying for years with their Internet2 concept, but so far they've not made much progress.

by Canada 4y ago

boltzmann-brain 4y ago

That's not really true. The ITU dictates what all sorts of telco equipment has to adhere to, and a bad actor could make it so that only compromised hardware (eg Russian or Chinese produced, with clear state involvement) can follow these standards. The IETF deals with a bunch of stuff, but that stuff runs _on top of_ what the ITU deals with. See here for more: https://news.ycombinator.com/item?id=33032452 https://news.ycombinator.com/item?id=33032452

tialaramex 4y ago
The ITU and IETF are very different. Although these are often lumped together as "Standards organisations" we should squint even at the word "Organisation" when it comes to the IETF. The IETF isn't an organisation - like the ITU or a Boy Scouts troop - it's an Activity, like dancing. Humans can participate in the IETF if they want, corporations cannot (corporations also can't dance) and since it isn't an organisation the IETF has no members and writes documents reflecting the consensus of participants. In this particular case an even more important difference is their approach to standardisation itself, the ITU produces "de jure" standards. The organisation produces documents which say how something shall be and that has the force of law. ITU members are sovereign entities, so they have police forces and armies, in principle they can choose to "enforce" the standards they've agreed, and only in practice do we end up relying on the fact that you might as well implement say I.762 as make up some other way to do it. The IETF produces documents how participants agreed things ought to work, but sometimes that's how they really work ("de facto" standards) and sometimes it isn't. For example the IETF published a document about how early browsers wanted Cookies to work, which briefly mentions how Netscape Navigator's first implementation did it, assuming this will quickly be superseded. It wasn't, the IETF RFC about cookies is only really useful today as either an aspiration or because it obliquely documents what you actually need to do, mimic Netscape (and guess which shortcuts work and which do not). The ISO and ITU approach works great where the main objective is to agree something and what it is exactly is secondary. For example the A-series paper standards from ISO are very nice, but if ISO had standardised what is now "US Letter" instead that would still be pretty useful even though it's an ugly shape without the nice properties of the A-series. However as you get into technical details this approach works much less and less well. In contrast the IETF approach works better for these details, because of the objective of consensus rather than an insistence on democracy, it will usually identify when there actually just isn't any agreement and standardise only what's actually agreed - the cookies example I mentioned is an example where this didn't work, but it's not that common. This has led to some conflict with "traditional" standards organisations which feel like they ought to just be able to write a document saying "No, it works like this" and have that take effect. Whether its ECMA, or the ITU, or even ISO this doesn't tend to go well for them but you can imagine it's frustrating even as they operate a liaison mechanism, to find that somehow the IETF is good at this while they can't comprehend how it even works.
wmf 4y ago
I would say the Internet has not run "on top of" telco networks for 10-20 years. Smart ISPs are running IP over MPLS over Ethernet over simple DWDM (OTN and even ROADMs are being dropped) which are mostly IETF and IEEE standards. When it comes to cellular, 3GPP appears to be upstream of ITU so we need to be vigilant against 3GPP getting corrupted but I haven't heard about that happening yet.
- fweimer 4y ago
  It depends. DSL and GPON are ITU-T standards. DOCSIS and Ethernet are not. 3GPP (mobile networks) does not seem to be ITU-T-dominated either, although they reuse some ITU processes. ASN.1 and X.509 are widely used in some Internet protocols and are ITU-T recommendations. I don't know what WAN looks like these days, but there seems to be some ITU standardization in the WDM area. So it's rather mixed. I'm not sure to what degree ITU-T standardization of obnoxious features matters in practice. Not everything gets implemented. A lot of dual-use features are present on ISP networks in Western democracies because they are required by law. Institutional safeguards generally keep users from government harm in these countries, and not lack of standardization of the more invasive surveillance technologies.
  - throw0101a 4y ago
    > It depends. DSL and GPON are ITU-T standards. The IEEE has PON stuff: * https://en.wikipedia.org/wiki/Ethernet_in_the_first_mile#Passive_optical_network https://en.wikipedia.org/wiki/Ethernet_in_the_first_mile#Pas... Though most PON deployments seem to be ITU-based… for historical/inertial/first-mover reasons?… but that can be changed by throwing money at the 'problem' (should it become one).
cryptonector 4y ago
Yeah, but the ITU-T gets no say on what happens at the IP layer and up (and also at some layers below IP, like MPLS). They want a say, and they've been trying for years with their Internet2 concept, but so far they've not made much progress.