Skip to content

Category: Networking

RTT-based vs. drop based congestion management

APNIC discusses different TCP congestion control algorithms, coming from Reno, going through CUBIC and Vegas, then introducing BBR (seems to be a variation on CoDel) and what they observed when running BBR in a network with other implementations.

TCP congestion control algorithms try to estimate the bandwidth limit of a multi-segment network path, where a stream crosses many routers. Each segment may have a different available capacity. Overloading the total path (that is, the thinnest subsegment of the path) will force packet drops by overloading the buffers of the router just in front of that thin segment. That in turn requires retransmits, which is inefficient and has nasty delays.

To make matters more complicated, the Internet is a dynamic environment and conditions can change during the lifetime of a connection.


The attack of the killer microseconds

In the Optane article I have been writing about how persistent bit-addressable memory will be changing things, and how network latencies may becoming a problem.

The ACM article Attack of the Killer Microseconds has another, more general take on the problem. It highlights how we are prepared in our machines to deal with very short delays such as nanoseconds, and how we are also prepared to deal with very long delays such as milliseconds. It’s the waits inbetween, the network latencies, sleep state wakeups and SSD access waits, that are too short to do something else and too long to busy wait in a Spinlock.

1 Comment

FOSDEM: The coming Radio Lockdown

The European Radio Equipment Directive requires all devices that are able to send and receive radio signals to be locked down. Without further specification of exceptions, which has not yet been done, this will affect all devices, including pure receivers such as GPS receivers and car radios, but also mobile phones and amateur radio operators and of course almost all Internet of Trash (IoT) devices.

Hardware manufacturers are required to “install technical measurements to protect the devices from being flashed with ‘non-compliant software'”.

The talk by Max Mehl is available on the FOSDEM site.


Dave Täht, LEDE and CoDel

Dave Täht

This is Dave Täht. Dave is working on LEDE. If you have been working with OpenWRT in the past, you should switch to LEDE. You should also give a lot of money to the Patreon of Dave.

Why should you be doing this?

Dave has been working on Networking Theory and Practice, and has been implementing CoDel. CoDel is one of the few innovations in basic networking, a queueing theory and algorithm that can make thick and fast internet pipes actually fast and efficient by managing buffer memory right.  That means that your Internet is fast, and feels fast, even if it is busy.

Getting CoDel right on Wifi is doubly hard because of interaction between the CoDel queues, Wifi media access control and Wifi encryption. But Dave has done that, and it’s part of LEDE.

So, you should

1 Comment

Netflix vs. IP v6

Netflix detects a tunnel

So in order to view Netflix, your network connection must be direct and not via a proxy or VPN tunnel.

Netflix, being somewhat modern, also advertises IP v6 services and AAAA DNS records so that your computer can find them.

On the other hand, many providers do not offer IP v6 natively, and hence require that customers who want non-legacy internet get it via – right – a network tunnel.

Which triggers the Netflix error message shown above.

Netflix knows that, but offers little support besides “Don’t use a tunnel, then”. Haha. So this article explains how to unfuck Networking for a local Linux or a Chromecast to make Netflix work again. Even if that just means to force it to fall back to l;egacy Internet instead.