3 ms·

> Again, that's not really how the internet has traditionally worked, and I'm not sure we really want content-based prioritization, outside of existing QoS stan

> Again, that's not really how the internet has traditionally worked, and I'm not sure we really want content-based prioritization, outside of existing QoS standards that aren't guaranteed across public internet connections. There's no reason QoS definitions couldn't be reviewed to re-prioritize specific IoT or autonomy protocols, but I don't see how the FCC should be making content decisions for consumers of internet bandwidth. I read this over and over but I don't understand. We have content-based prioritization with QoS and if the Title 1 classification went ahead wouldn't the FCC be in the position of regulating those QoS standards which as you point out, are managed privately at the moment. Maybe im confused but I believed based on the arguments made by the FCC under Obama that the FCC was looking to regulate QoS rules set by ISP's and potentially other network providers.

wtallis 9y ago
> We have content-based prioritization with QoS and if the Title 1 classification went ahead wouldn't the FCC be in the position of regulating those QoS standards which as you point out, are managed privately at the moment. The Internet as it currently exists has DiffServ markings, the meaning of which is not clearly defined, and the markings themselves are not reliably preserved end-to-end. In practice they are not usable for connections passing through multiple networks with different owners. The markings could be set by the endpoints to request what amounts to content-based prioritization, but the markings aren't tamper-proof. There are no other QoS technologies that are available and supported on the public Internet. ISPs may be doing things behind the curtain, but there's no transparency or control available to their users. Recent years have brought a renaissance to the QoS technologies available for users to deploy in their own gateway routers and new congestion control algorithms to use on their own endpoints, but the former isn't being adopted by ISPs and the latter can't be. However, those new QoS technologies have shown that need for DiffServ-style QoS on the public Internet is largely fictitious, driven by ignorance of the advances in QoS methods since the early 1990s. It is now clear that outlawing ISPs' use of discriminatory DiffServ or port-based QoS rules can be justified on purely technical grounds, without even bringing up the business conflicts of interest involved. Content-based prioritization rules don't hold the record for best-performing QoS system, even when the rulesets are carefully curated.

wtallis 9y ago

> We have content-based prioritization with QoS and if the Title 1 classification went ahead wouldn't the FCC be in the position of regulating those QoS standards which as you point out, are managed privately at the moment. The Internet as it currently exists has DiffServ markings, the meaning of which is not clearly defined, and the markings themselves are not reliably preserved end-to-end. In practice they are not usable for connections passing through multiple networks with different owners. The markings could be set by the endpoints to request what amounts to content-based prioritization, but the markings aren't tamper-proof. There are no other QoS technologies that are available and supported on the public Internet. ISPs may be doing things behind the curtain, but there's no transparency or control available to their users. Recent years have brought a renaissance to the QoS technologies available for users to deploy in their own gateway routers and new congestion control algorithms to use on their own endpoints, but the former isn't being adopted by ISPs and the latter can't be. However, those new QoS technologies have shown that need for DiffServ-style QoS on the public Internet is largely fictitious, driven by ignorance of the advances in QoS methods since the early 1990s. It is now clear that outlawing ISPs' use of discriminatory DiffServ or port-based QoS rules can be justified on purely technical grounds, without even bringing up the business conflicts of interest involved. Content-based prioritization rules don't hold the record for best-performing QoS system, even when the rulesets are carefully curated.