Some Internet Outages Predicted For the Coming Month as '768k Day' Approaches (zdnet.com) 65
An internet milestone known as "768k Day" is getting closer and some network administrators are shaking in their boots fearing downtime caused by outdated network equipment. From a report: The fear is justified, and many companies have taken precautions to update old routers, but some cascading failures are still predicted. The term 768k Day comes from the original mother of all internet outages known as 512k Day. [...] Many legacy routers received emergency firmware patches that allowed network admins to set a higher threshold for the size of the memory allocated to handle the global BGP routing table. Most network administrators followed documentation provided at the time and set the new upper limit at 768,000 -- aka 768k.
CIDR Report, a website that keeps track of the global BGP routing table, puts the size of this file at 773,480 entries; however, their version of the table isn't official and contains some duplicates. A Twitter bot named BGP4-Table, which has also been tracking the size of the global BGP routing table in anticipation of 768K Day, puts the actual size of the file at 767,392, just a hair away from overflowing. ZDNet spoke today with Aaron A. Glenn, a networking engineer with AAGICo Berlin, and Jim Troutman, Director at the Northern New England Neutral Internet Exchange (NNENIX). Both estimate 768K Day happening within the next month. But unlike many network admins, they don't expect the event to cause internet-wide outages like in 2014. However, both Glenn and Troutman expect some companies and smaller, local ISPs to be affected. "I would be mildly surprised if there was any interruption or outage at any real scale," Glenn told ZDNet.
CIDR Report, a website that keeps track of the global BGP routing table, puts the size of this file at 773,480 entries; however, their version of the table isn't official and contains some duplicates. A Twitter bot named BGP4-Table, which has also been tracking the size of the global BGP routing table in anticipation of 768K Day, puts the actual size of the file at 767,392, just a hair away from overflowing. ZDNet spoke today with Aaron A. Glenn, a networking engineer with AAGICo Berlin, and Jim Troutman, Director at the Northern New England Neutral Internet Exchange (NNENIX). Both estimate 768K Day happening within the next month. But unlike many network admins, they don't expect the event to cause internet-wide outages like in 2014. However, both Glenn and Troutman expect some companies and smaller, local ISPs to be affected. "I would be mildly surprised if there was any interruption or outage at any real scale," Glenn told ZDNet.
IPv4 fragmentation (Score:2)
Re: (Score:2, Insightful)
Networking has been retardedly using 1000 B for 1 KB forever.
It comes from the fact that in ye olden days, we used analog devices that operated in baudrate and symbols per second.
The morons never made the correction from baud (often denoted as b) to bit (b) or byte (B) when dealing with anything on the order of 1000 (or higher).
It's so bad the the IEEE entrenched it as official bullshit. We don't really have gigabit networking, we have 1000 megabit networking (which in turn...).
1 KB = 1024 B. Anyone who s
Re: (Score:2)
But that's the thing; KB was never meant to be a SI unit or confused with one, but a distinct compound unit. Hence for years there was an argument about whether the k should be capitalised. Yes, it should, if you're meaning 1024 of the little blighters. Of course, mega being a capital M in the first place didn't help matters...
Re: (Score:2)
So when will the SI people accept the proper term for 1,000 Kilograms ie Megagram
Stop using the ton(ne)
Re:maybe they should use iphones for routing (Score:5, Interesting)
Re:maybe they should use iphones for routing (Score:5, Informative)
An IPv4 address is 4 bytes, so even if you were to write software which could process two addresses (64 bits) at once, your iPhone would be roughly 400,000 times slower than these devices. I suppose that's better than being 800,000 times slower due to running 32-bit instructions. But the 2x speedup from using 64-bit instructions instead of 32-bit is infinitesimal compared to the 400,000x speedup from using TCAM instead of RAM.
Re:maybe they should use iphones for routing (Score:4, Funny)
Re: (Score:2)
For the cesspool of shit that Slashdot has become I still come here for posts like this. Thanks. Quite informative and I learnt something new.
32 bit, 64 bit, 8 million bit. 500,000X faster (Score:5, Informative)
The Apple A8 SOC can fetch 64 bits from memory in a clock cycle, and compare them to a known value in another clock cycle. This is a binary compare, 1 or 0.
The Cisco TCAM can compare 8 MILLION bits in a single clock cycle. So it's 250,000 times faster.
Also the router is doing a ternary compare; 1, 0, or don't care. That's necessary because a route may be for 67.53.1.1/19, we don't care about the leftmost 13 bits. The ARM would require at least two additional clock cycles to shift out the don't care bits, making it 500,000 times slower.
Re: (Score:2)
and these jokers can't make routers that cost tens of thousands of $$$ that hold more than 768k of routes?
The only joker here is you.. except you arent joking.. this is news for nerds, its not supposed to be geeks pretending to be nerds
FUD (Score:5, Interesting)
Re:FUD (Score:4, Informative)
As engineers around the world rushed to fix the problem, one internet traffic monitor reported that the number of routes suffering outages on Tuesday had doubled to 12,600. On a typical day outages normally affect 6,000 routes.
So, as you say, not really sky falling.
Re: (Score:2)
Things like this end up to be boring, small problems because people worry about them and talk about them and outreach causes people to fix the problem before it happens or in a short time after it happens.
If 512k route limits were ignored it would have been a much bigger problem. Or Y2K. Or whatever.
(768k will be even less of a problem, because a lot of the affected routers have since been replaced with hardware with higher limits in the last 5 years).
Re:FUD (Score:5, Informative)
Re: (Score:3)
:) I never understand people who assert this kind of thing.
I mean, is the line of argument "Hey, with extensive mitigation effort, it was only a moderate problem--- so we shouldn't even have fixed anything!" ?
Re: (Score:2)
The nice upshot of it all was the stuff that couldn't be upgraded was just outright replaced
That's not a nice upshot here. Becoming wasteful with hardware is one of the things that is holding back actually solving the problem, which is that IP addressing was never designed to be this fragmented.
I kinda wish better hardware wasn't available. Maybe then we'd be having this conversation on IPv6 instead of on an old line which I had to pay extra for just so I could get an IP address that was wasn't stuck behind a CGNAT device.
Re: (Score:1)
Oh really now? How about this [slashdot.org]?
Bill Gates (Score:1)
How many people actually heard of 512k day before this post?
Bill Gates has, and I think you know his feelings on whether it should be enough for anybody.
Re: (Score:2)
You're falling into the same trap as the folks who said nothing happened on Y2K so it wasnt'a big deal. It was only not a big deal because of the work preparing for it in advance.
Re: (Score:2)
Y2K also wasn't a big deal. Mostly 'cause a lot of good people worked 24/7 to make sure it ain't gonna be a big deal.
It's kinda disheartening, ya know? When you save the world (ok, with a bit less melodrama, when you allow every Billy-Bob out there to access his porn uninterrupted), you feel like you're on top of the world. Until you hear Billy-Bob say "well, crying wolf again?"
768K should be enough for anyone (Score:3, Funny)
Re: (Score:1)
Multiple people remember him making similar statements. At the very least, there's a lot of evidence he was surprised by how fast software vendors "used up" RAM as the upper practical limit grew.
Re:768K should be enough for anyone (Score:4, Funny)