6 ms·
I've wondered about the fact that (AFAIK) Tor routes only TCP: 1. Is that (still) correct? 2. Can't web pages include non-TCP traffic, and if so, is it routed
by wolverine876 5y ago
I've wondered about the fact that (AFAIK) Tor routes only TCP:
1. Is that (still) correct?
2. Can't web pages include non-TCP traffic, and if so, is it routed via Tor? For example, doesn't some some streaming media use UDP?
3. QUIC doesn't use TCP (deliberately, I think). Won't that affect Tor's long-term viability if everyone eventually moves to QUIC?
- zamadatix 5y ago1. For now 2. That kind of traffic mostly died out with plugins like Flash, streaming media today is usually encapsulated in small frequent chunks over the same https channels as the rest of the page are delivered as that's what's available in JS/native. WebRTC reintroduced some UDP stuff and can be used for streaming but is still mostly used for peer to peer calls. 3. https://gitweb.torproject.org/tor-browser-spec.git/plain/position-papers/HTTP3/HTTP3.pdf https://gitweb.torproject.org/tor-browser-spec.git/plain/pos...
- wolverine876 5y agoThank you! 2. So there is some non-TCP traffic. What happens when you load that page in Tor Browser, for example? Does it leak back to your clear Internet connection? Is it simply dropped? This seems like a critical issue. 3. Thanks. Do you know when that was written? To save others clicking the link and finding the applicable section, I'll paste it below. Designing and building your own protocol for Internet transport, compatible with the entire net and performing competitively enough to be usable, sounds like quite a project for a small organization. Note that Google didn't do that; they used UDP for QUIC. 7 Tor Network Compatibility Concerns Our final area of concern is continued compatibility of the Tor network with future versions of the HTTP proto- col. It is our understanding that there is a desire for future versions of HTTP to move to a UDP transport layer so that reliability, congestion control, and client mobility will be more directly under control of the client user agent. At present, the Tor Network is only capable of carrying TCP traffic. While it will be possible to support the transit of UDP datagrams using our existing TCP overlay network without significant anonymity risks within a year’s time or sooner, it is unlikely that this level of support will be sufficient to warrant the use of a finely-tuned UDP version of HTTP rather than a TCP variant. Long term, our goal is to transition the entire Tor network to our own datagram protocol with custom con- gestion and flow control to better support both native datagram transport and end-to-end flow control. However, additional research is still needed to examine the anonymity implications associated with this transition[12]. Our present estimate is that a full network transition to UDP is at least five years away. We are also concerned that even after a full network transition to a datagram transport, it is likely that the congestion, flow, and reliability control of a UDP version of HTTP may still end up performing poorly over higher-latency overlay networks such as ours. For these reasons, we are especially interested in ensuring that overlay networks are taken into account in the design of any UDP-based future versions of HTTP, and also prefer to retain the ability to use future HTTP versions over TCP, should the UDP implementations prove sub-optimal for our use case.
- scandinavian 5y ago2. TOR browser is built without webrtc support. There is no other way afaik, as firefox doesn't have a socket API for extensions anymore.
- olddb 5y ago> streaming media today is usually encapsulated in small frequent chunks over the same https channels as the rest of the page are delivered as that's what's available in JS/native $> --- HLS ---<3
- dillondoyle 5y agoIt's easy to get a webrtc fingerprint just using a public stun server, maybe people smarter can deploy their own. I've used it in our ad tracking js. I'm not sure if Tor Browser turns off by default, searching found this one ticket which suggest that default flag but maybe it's not implemented out of the box. [1] https://privacycheck.sec.lrz.de/active/fp_wrtc/fp_webrtc.html https://privacycheck.sec.lrz.de/active/fp_wrtc/fp_webrtc.htm... [2] https://gitlab.torproject.org/legacy/trac/-/issues/8178 https://gitlab.torproject.org/legacy/trac/-/issues/8178
- zamadatix 5y ago> I'm not sure if Tor Browser turns off by default, searching found this one ticket which suggest that default flag but maybe it's not implemented out of the box. It does and it is removed at compile time since that ticket was closed (i.e. that was a build flag not a runtime flag).
- dillondoyle 5y agoThanks for explaining!
- georgyo 5y agoQUIC and HTTP3 are great technologies, but they are never likely to become the only protocol a service supports. For one thing, convection to a website via one of those protocols first, and then a header informs the client that it can reconnect via QUIC/HTTP3. IE they have to have a working http 1 or 2 webserver first. UDP is disallowed in many many places, and many ISPs treat UDP as hostile and rate limit it. In the places it works, it provides some benefits. But we're unlikely to see it take over as the sole protocol any time soon.
- Seirdy 5y ago> For one thing, convection to a website via one of those protocols first, and then a header informs the client that it can reconnect via QUIC/HTTP3. Nit: new SVCB DNS records can serve the same purpose as Alt-Svc HTTP headers before the initial request, so the first request to a server is HTTP/3. But yeah, HTTP/1.1 isn't going away (and shouldn't go away) for many reasons.
- wolverine876 5y ago> we're unlikely to see it [QUIC] take over as the sole protocol any time soon Agreed, but I'm not talking about soon. I mean the long term. Even FTP has been deprecated.
- georgyo 5y agoI think when that starts to become a reality solutions will be discovered. But since it is provably a non-issue today because it requires upgrading from TCP, it's going to be low priority.
- skissane 5y ago> Even FTP has been deprecated Not everywhere. FTP-over-TLS is secure, standardised (RFC4217 as updated by RFC8996), and in some environments is still preferred to SFTP, particularly mainframe and minicomputer environments. FTP, due to its age, has a lot of "legacy" features which mean it can work better with non-POSIX filesystems used on mainframe and minicomputer systems than SFTP can. In principle you could add extensions to SFTP to improve its support for non-POSIX filesystems, but why bother when FTP already has very well-established support for that? Another area in which FTP is still preferred is transfer of very large (multi-terabyte) scientific datasets. GridFTP has defined FTP extensions which permit these transfers, including encryption and striping of files across multiple connections and servers (so multiple servers can cooperate to simultaneously transfer different portions of an extremely large file). SFTP has no advantage for this application, and why bother redefining those extensions over SFTP when they work perfectly well over FTP? The main competitor to GridFTP is not SFTP, but rather proprietary solutions such as IBM Aspera. GridFTP actually supports SSH as a transport, but even then the file transfer protocol is based on FTP not the binary SFTP protocol. Similar comments apply to TELNET. TELNET-over-TLS is secure, and still preferred in some IBM environments, because there are established protocols for passing 3270 and 5250 block mode terminal data streams over TELNET. Again, no reason in principle why you couldn't define similar protocol extensions for SSH, but why bother when TELNET works perfectly well for this application? And if you really want to use SSH instead of TLS as a transport/security layer, nothing stops you from tunnelling TELNET over SSH.
- pabs3 5y agoI wonder how many networks allow UDP (and thus QUIC) to pass, seems like there will be lots of places where QUIC can't be used.