Upgrading Fiber To The Home to terabit speeds

▼ Last week, Jaap van Till asked me if BGP would be capable of supporting the terabit class interconnectivity that he foresees we’ll need in the future, possibly due to the rise of artificial intelligence. He explains his reasoning in the blog post What Link speeds will we need for AI, where he quotes VAN TILL’s CONJECTURE:

The network connection Wide Area access speed will grow in time until it matches the internal device BUS speed of the more and more complex processors and datastores.

And then concludes that 14 Tbps external links will be required in 2039. Today I can get 4 Gbps where I live. So that means a 70% speed increase per year.

Let’s first get that BGP question out of the way: I see no problems. 25 years ago I ran BGP over 64 and 128 kbps links without trouble. Six orders of magnitude later, BGP is still fine, and there is no reason to believe that even faster speeds will be a problem, just as long as the packet loss rates remain minimal.

But what would terabit class network connectivity at home look like?

Actually, I think we have all the parts to build this today. With Wavelength Devision Multiplexing (WDM), it’s possible to transmit multiple data streams through a single fiber by using slightly different wavelengths/frequencies of infrared laser light. Coarse WDM (CWDM) is relatively cheap and appropriate over shorter distances, with 18 wavelengths standardized over high performance fiber. (Fewer over most existing fiber.) For long distances, dense WDM (DWDM) can use as many as 160 wavelengths over a single fiber pair.

Bandwidth per wavelength is now 100 or 200 Gbps, and expected to increase in the future. So anything between say 10 x 100 Gbps = 1 Tbps and the 20 Tbps used by modern seacables should be possible. The catch is of course the cost.

The difficulty is with the transmitting side, as this requires a tuned laser per wavelength. On the receiving side, the wavelengths can be split using a prism and hit a set of wideband receivers. As someone who is definitely not in the business of building this equipment, it seems to me that a system with one or a small number of transmitters, a passive optical bus, and a large(r) set of receivers is definitely something that could enjoy radical performance vs cost improvements over time. And it fits perfectly with the most efficient / high speed way to connect homes to the internet that we have today: PON (passive optical network). So just add additional wavelengths to existing PON installations to gain more bandwidth in the downstream direction.

However, now we have a new challenge: TCP/IP is not a good fit for sending the massive data streams that would make good use of such a network. The problem is that TCP tries to adjust its end-to-end data transmission rate to available bandwidth. This means it needs to wait for acknowledgments from the receiving side to know it can increase its transmission rate, maintain it, or slow the transmission rate down. Downloading 100 MB over a 1 Tbps link takes less than a millisecond. But even over PON, the round-trip-time is a millisecond or two. This means that the bottleneck is the number of round trip TCP requires to reach that full terabit speed. Even if that’s an extremely unrealistic 10 RTTs, that means the total transmission time is now 11 ms, effectively only using a tenth of the available bandwidth.

So we need to overhaul TCP/IP for the super high speed stuff and instead use something more like circuit switching / time division multiplexing / token passing. Yes, everything old is new again! So for instance reserve ten 100 μs timeslots and transmit ten 10 MB “megapackets”.

So I think all of this is highly doable!

Well, there is the slight challenge of how to pipe all that bandwidth into your laptop without connecting/disconnecting that fiber all the time. Maybe use eight Thunderbolt 5 interfaces in parallel to reach 960 Gbps?

Permalink - posted 2024-04-09

BGP handling of obscure errors

▼ I read Ben Cartwright Cox' (extensive) blog post Grave flaws in BGP Error handling and then saw his talk about the same topic at NLNOG on Youtube.

The necessary background

In addition to the "well-known" BGP path attributes that we all know (because the RFC says we must) and love (because they make the internet work), it's also possible define new attributes to provide new functionality. These can be "transitive" attributes, which means that a BGP router that doesn't recognize them propagates them to its BGP neighbors unchanged.

The ability to create new optional transitive attributes has allowed us to run BGP version 4 for three decades without having to bump the version number because we had to make backward-incompatible changes that would make adoption all but impossible.

For instance, 32-bit autonomous system numbers were added as the 16-bit BGP AS numbers started to run out. In addition to the well-known (mandatory) 16-bit AS path, an optional 32-bit AS path was added. If a router in the middle didn't understand the 32-bit AS path, it would update the 16-bit AS path and propagate the 32-bit AS path unchanged.

The next 32-bit capable BGP router can then add back the AS numbers from the 16-bit path that are missing from the 32-bit path, and 32-bit AS numbers work even if routers in the middle don't understand them. (They just see "23456".)

Of course you can read all about BGP attributes in my book Internet Routing with BGP. It's even in the sample chapters! (Page 15.)

The error handling issue

In Ben's blog post, he talks about a Brazilian network included a malformed version of a still experimental attribute. All the big routers in the core of the internet don't run experiments, so they just saw an attribute they didn't recognize, and propagated it as per the transitive setting. Eventually BGP updates with the broken attribute arrived at routers that did understand the attribute, but saw that it was broken.

So as per the original BGP spec, they tore down the BGP session towards the router that sent them the broken attribute. And then, after a short delay, tried to set up a new BGP session towards that neighboring router. Only to encounter the same error again and tearing down the BGP session again. And so on.

Which is probably not wat you want. Which is nicely explained in RFC 7606, published in 2015, which suggests to treat such errors as if the neighboring router had asked to withdraw the route containing the offending path attribute. So if a neighbor tells me prefixes 10.0.0.0/8, 172.16.0.0/12 and 192.168.0.0/16 are reachable through them, and 172.16.0.0/12 has a broken attribute, I just act as if my neighbor had told me that 172.16.0.0/12 is not reachable through them. But I don't bring down the BGP session so 10.0.0.0/8 and 192.168.0.0/16 remain reachable through the neighbor in question.

Ben seems to be rather annoyed that many router vendors don't implement the RFC 7606 behavior, implement it but don't enable it by default, and/or don't have a bug bounty program to reward security researchers for pointing out these deficiencies. He spent a good amount of time evaluating different implementations and then "fuzzing" attributes to see what would happen, So that's somewhat understandable. Here is his score card from his presentation slides:

My take

I agree that the RFC 7606 handling by default is what you want. I also agree that changing a default here, something router vendors loathe to do, shouldn't be problematic.

However, these are pretty obscure errors. This is not an internet extinction level issue.

For my own network, I would strongly prefer a mechanism to turn off handling of these often rather frivolous new attributes. Both to avoid being bitten by buggy implementations elsewhere, but also to avoid inflating BGP messages. As BGP updates propagate, the AS paths (the 16- and 32-bit versions) increase in length, so an update that was just under the limit at some point will exceed the maximum size of 4096 bytes at some point, and then definitely bad things will happen.

However, it's important that new transitive attributes aren't filtered out wholesale, as that would make it impossible to add new features to BGP. I'm not sure if there is a workable way to put a stop to frivolous BGP path attributes being injected into the global routing system while at the same time not robbing BGP of its forward compatibility with future new innovations.

Permalink - posted 2023-10-02

Should the datacenter be in the middle?

▼ The other day, I landed on this article: In Focus: Subsea Network Architecture: IXPs. The article takes some time to arrive at the point that undersea internet exchanges would be a good idea. The most eyecatching part is a variation on this image:

As the article starts out discussing how datacenters have been moving away from large cities to take advantage of opportunities such as space, cheap energy and easier cooling, this image seems to suggest that these blue dots in are good locations for datacenters and/or internet exchanges in general. And that's definitely not the point of the paper that the image is from.

That paper is very specifically about the best locations to place servers for high speed algorithmic trading on multiple markets some distance away from each other. This immediately explains why there is nothing around the western US: there are simply no stock exchanges / markets there (the red dots in the image).

The math looks more complicated, but presumably, in these cases it helps when the servers executing the trading algorithms are in the middle between the "users", rather than close to one and further from the other(s).

If you need data from two places far away from each other, then it's better when each is 25 milliseconds away, as you can then complete your action in 25 ms plus however long it takes to do your own processing. If you're close to one so it's 0 ms for one data source and 50 ms for the other, then the entire action takes at least 50 ms.

But is that a common situation?

In general, you can just copy the data beforehand. So this only applies if you're using "live" data from two or more locations. Videoconferencing with a number of participants could be an example, where a server receives the video from all the participants, mixes it into a single feed and then sends that single feed out to all the participants. If the server is in the middle, this limits the maximum delay. I guess that could be somewhat helpful. But to the degree that it makes sense to have datacenters in the middle of the ocean? I'm not convinced.

Permalink - posted 2023-09-07