Comment by stek29

8 months ago

The post implies that ISPs are doing the throttling — but in fact, ISPs have little control over it lately. This kind of tampering is mostly done on TSPU devices (Technical Measures for Threat Counteraction), which is fully controlled by Roskomnadzor. ISPs have little control over it — the system is highly centralized, and Roskomnadzor can carry out sophisticated and targeted actions. For example, they're able to swiftly block some services and networks in specific regions and/or on specific types of connections (cellular vs wired, or even exact ISPs), or to perform

For example, when YouTube blocking first began, TSPUs were only throttling it on wired networks. This created an unfair advantage for cellular ISPs, prompting many people to cancel their wired internet contracts and switch to mobile networks.

Since there were no legal grounds for the block, and the government even denied the blocks, trying to blame the throttling on Google, some ISPs just started circumventing the blocks, effectively counteracting the TSPU blocks.

They soon had to stop doing that because of warrants and threats from Roskomnadzor — but it clearly illustrates that ISPs have virtually no control over internet blocking — in either technical or legal terms.

As for Cloudflare specifically — there’s more to the story, and I’m disappointed the post didn’t include more details.

Roskomnadzor is blocking TLS 1.3 ECH — and they are doing it for Cloudflare specifically, forcing Cloudflare users to disable ECH on their domains if they want their websites to be accessible in Russia when using up-to-date browsers.

This also means that, for the website using Cloudflare to be accessible, it needs plaintext SNI. Cloudflare also does not allow domain fronting.

This has allowed Roskomnadzor to implement selective allowlist-based blocking on Cloudflare networks, which is a new step in Russian internet censorship. It feels like Cloudflare could've done a more in-depth analysis of selective blockings, since they probably can see whether there are statistically noticeable differences in metrics for different domains.

References (some are in Russian):

On some regions having some services like Telegram blocked for months:

- https://ntc.party/t/в-дагестане-отключили-зарубежный-интерне...

- https://www.forbes.ru/tekhnologii/532303-v-dagestane-i-cecne...

- https://www.vedomosti.ru/technology/articles/2025/05/31/1114...

On ISPs getting threats from Roskomnadzor for un-throttling YouTube:

- https://www.cnews.ru/news/top/2024-08-29_mgnovennaya_reaktsi...

- https://www.forbes.ru/tekhnologii/520476-uskorausim-youtube-...

On cloudflare-ech being blocked (since November 2024):

- https://github.com/net4people/bbs/issues/417

- https://ntc.party/t/обсуждение-блокировка-замедление-ech-clo...

on allowlists:

- https://ntc.party/t/09062025-информация-по-блокировке-cloudf...