dtaht

12 ms·

91.
by dtaht 3y ago
The bottleneck is mostly on the read path,even leveraging xdp heavily. dpdk/vpp would be faster. In the case however of an ISP 10k subscribers/network segment at 25Gbit is more than enough and busies 40% of 20 cores, on libreqos.
92.
25 years since the first embedded wireless Linux routers(rage.net)
2 points by dtaht 3y ago|1 comments
93.
by dtaht 3y ago
I last wrote about the wild ride wifi took us all on back in 2010. But I had a huge flashback just now, of sitting on my roof, in the rain, in 1998,after re-aligning our antenna - and wondering where would wireless go, and how much fun we h
94.
by dtaht 3y ago
I think FQ helps everywhere, and FQ + aqm is needed at every fast->slow transition on the network. Ideally the core is over provisioned... except where it is not.
95.
by dtaht 3y ago
We designed cake to be implementable in hardware if anyone could be convinced it was useful: https://blog.cerowrt.org/post/juniper/ - I thought this test series rather compelling. 1) The data paths MUST prepend a
96.
by dtaht 3y ago
Wish more would look harder there at the actual delays being observed and how bad congestion collapse gets at +250ms of it. https://blog.cerowrt.org/post/juniper/
97.
by dtaht 3y ago
Pushing CAKE to 10Gbit/xeon core in libreqos currently (the bottleneck is on the read path actually). Or, 10k ISP subscribers per $2k xeon box.
98.
by dtaht 3y ago
Nice to see you mark! and that you are using CAKE. Cake has a feature that helps on regulating multiple flows to a single destination: per host/per flow FQ. In the middle of a network this only works well with IPv6, along the edge it h
99.
by dtaht 3y ago
We are making significant progress with CAKE & fq_codel in the WISP & and fiber markets as a middlebox. See Preseem, Bequant, and LibreQos.io. Still best to get these native on the CPE as Mikrotik and many many others have one. I wo
100.
by dtaht 3y ago
I miss bump: https://en.wikipedia.org/wiki/Bump_(application) I rather hate the gross over-characterizations of SV as being VC-led (or destroyed). There are plenty of old fashioned network hippies still there, very con
101.
by dtaht 3y ago
25 years ago today, the first of Microsoft's internal memoranda regarding how it was going to deal with the "Linux threat" were leaked to the press. Microsoft at the time was coping with a DoJ lawsuit. Linux itself, its commu
102.
25 Years Since the Halloween Microsoft vs. Linux Memos(catb.org)
7 points by dtaht 3y ago|1 comments
103.
by dtaht 3y ago
I use sqm-scripts for a variety of purposes to insert delay or other variables like aqm and fq technologies. I also wrote a now somewhat out of date cautionary note about all the pitfalls you can land into here: https://www.buffe
104.
by dtaht 3y ago
Have you ever read any Buckminster Fuller? "Spaceship Earth" had a huge influence on me.
105.
by dtaht 3y ago
I have just taken the time to point out a few things he did more quietly, later in life, with enormous impact.
106.
by dtaht 3y ago
There is an even sadder backstory to the whole LF thing, and the time it got created and sucked the air out of the room. I wish we had found a way to get this idea off the ground, and established the academic and early response institution
107.
by dtaht 3y ago
One of my great sadnesses about the long term support of FOSS was that the great redhat experiment of letting open source contributors into the IPO´s strike price, has never been repeated. For a tiny amount of re-investment into the commo
108.
by dtaht 3y ago
The core infrastructure initiative sucked a lot of air into their room, and accomplished almost nothing.
109.
by dtaht 3y ago
I strongly disagree that esr´s books are irrelevant. You had to be there, in 1998, I guess, facing the prospect of being fired for working on free software, as many were, or facing down the might of the legal apparatus of microsoft, and adm
110.
by dtaht 3y ago
I should totally take that down. Nobody cared. I got, oh, perhaps 70 dollars a month more into my patreon via this idea... ... but I like to think, that the awareness of the problem itself has propagated all the way up to the highest levels
111.
by dtaht 3y ago
I just heard about that recently! I am encouraged! I hope they put some dough under OpenWrt. NLNET has also been a consistent and easy to apply to funder.
112.
by dtaht 3y ago
He never cashed out. He got nothing. Just the royalties on his books and the occasional consulting contract.
113.
by dtaht 3y ago
His idea was published in the last linux journal, and generated in part, in sympathy for my state of health at the time. Since then there has been a lot of decent research on the "load" and emerging government support for critical
114.
by dtaht 3y ago
I am surprised to see this show up again. The idea was ahead of its time, and sank without a trace. Since then, (and during) multiple orgs have attempted to also put some sort of floor below various maintainers. I see these finally beginnin
115.
by dtaht 3y ago
Moving forward I hope policy decisions are made with an eye towards truly making the internet faster. https://blog.cloudflare.com/making-home-internet-faster/
116.
by dtaht 3y ago
All the that is needed for evil to triumph is for good men to do nothing.
117.
by dtaht 3y ago
That is a beautiful thread as to what was the original source of the network neutrality bittorrent vs voip bufferbloat blowup. https://twitter.com/jlivingood/status/1707078242857849244 With tons of links to all th
118.
Net Neutrality´s origins in bufferbloat, BitTorrent vs. VoIP(twitter.com)
3 points by dtaht 3y ago|3 comments
119.
by dtaht 3y ago
If only more political folk understood how impossible it was in 2005 to achieve "common carriage" for two very different protocols - bulk torrents, vs VOIP - and how solved that problem is now, technically (fq_codel, CAKE, PIE, BB
120.
by dtaht 3y ago
Thank you for leveraging it. but: Eric Dumazet came up with the flow queuing fast/slow queue concept in a single saturday afternoon in his hammock (after admittedly 9 months of hard work on every other fq and aqm algorithm), slammed it

More ›