Tag Archive for Y2K

Should I Care About 768k Day?

Why Should I Care About 768k Day?If you are of a certain age, you remember Y2K. While I was not rewriting COBOL programs, I played my part. I spent the last half of 1999 scheduling after-hours downtime to update Compaq 1900 and 2500 servers with BIOS updates on a floppy disk. Hoping and praying the servers would come back up after the floppy disk stopped grinding. As I recall only two Compaq Proliant 2500‘s failed the BIOS upgrade and only one was DOA.

All the fun of Y2K was because memory space was too small to accommodate the fancy new year 2000 without thinking it was 1900. Now a similar memory size problem could cause internet disruptions very soon. The problem is called 768k Day.

768k Day is when the size of the global BGP routing table is expected to exceed 768,000 entries. Anthony Spadafora at TechRadar explains that on August 12, 2014, a similar problem, occurred after Verizon (VZ) advertised 15,000 new BGP routes to the internet. Verizon’s actions caused the global BGP routing table, a file that holds the IPv4 addresses of all known internet-connected networks, to exceed 512,000 causing the 512K Day crisis.

Over flowingThe TechRadar article explains that in 2014, ISPs and others had configured the size of the memory for their router TCAMs (ternary content-addressable memory) for a limit of 512K route entries and some older routers suffered memory overflows which led their CPUs to crash. These crashes created significant packet loss and traffic outages across the internet with even large provider networks being affected. ZDNet says companies like Microsoft, eBay, BT, Comcast, AT&T, Sprint, and Verizon, were all impacted by 512K day

Engineers and network administrators rushed to apply emergency firmware patches to set a new upper limit which in many cases was 768k entries. The seeds of the 2019 768k  crisis were sown.

preventative maintenanceMr. Spadafora speculates that in 2019 most of the large providers who felt 513K day’s impact have likely updated and maintained their infrastructures reasonably well which could lead to fewer outages. He says that there are still ‘soft spots’ smaller ISPs, data centers, and other providers who are part of the Internet’s fabric where maintenance on legacy routers and network equipment can be neglected or missed more easily.

These are the places that most likely see some issues or outages due to 768k Day. These outages will create significant packet loss and traffic outages that could have a ripple effect and sweep upstream and affect larger provider networks. Alex Henthorn-Iwane at network intelligence firm ThousandEyes writes,Given the sheer size and unregulated nature of the Internet, it’s fair to say that things will be missed.

rb-

To prepare for any potential disruptions, it is a good idea to perform some preventative maintenance on any routers that receive full internet routes. Jim Troutman, Director at the Northern New England Neutral Internet Exchange (NNENIX) told ZDNet,

The 768k IPv4 route limit is only a problem if you are taking ALL routes. If you discard or don’t accept /24 routes, that eliminates half the total BGP table size.

There is still a little time left before 768K day, at 2019-06-21 16:00 UTC 06/21/2019 the Regional Internet Registry for Europe, the Middle East, and parts of Central Asia (RIPE) reports that 86.9% of the IPv4 BGP tables they monitor are below 768K. Click here for current results

What is the big deal? Network intelligence firm ThousandEyes points out that there are many outage events that happen every day, especially on the fringes of the Internet. The number of garden variety outages could get amplified because of 768k day-related issues over the next few weeks.

Aaron A. Glenn, a networking engineer with AAGICo Berlin told ZDNet,

Cisco 6509The Cisco 6500/7600 product line was extremely popular for an exceptionally long time in many, many places,” so don’t be surprised if some networks go offline because they forgot about 768k Day and didn’t prepare.

Related articles

 

Ralph Bach has been in IT long enough to know better and has blogged from his Bach Seat about IT, careers, and anything else that catches his attention since 2005. You can follow him on LinkedInFacebook, and Twitter. Email the Bach Seat here.

70s Glitch Could Hit Every Computer On Earth

70s Glitch Could Hit Every Computer On The PlanetRebecca Borison at the BusinessInsider asks who remembers the 1999 panic about the Y2K crisis. In 1999, Y2K looked as if it might derail modern life when computers because the glitch would reset computers to Jan 1. 1900, rather than Jan. 1, 2000, because computers only used two digits to represent a year in their internal clocks.

déjà vu all over againNow it déjà vu all over again, BI reports there’s a new, even bigger global software coding fiasco looming.  A huge amount of computer software could fail around the year 2038 because of issues with the way the code that runs them measures time.

Once again, just like with Y2K every single piece of software and computer code on the planet must now be checked and updated again. That is not a trivial task according to the author. In 2000, we bypassed the Y2K problem by recoding the software explains Ms. Borison. All the software — a fantastically laborious retrospective global software patch.

Disruption to the tech industry

Y2K problemAlthough Y2K was not a disaster, it was a massive disruption to the tech industry at the time. Virtually every company on the planet running any type of software had to find their specific Y2K issue and hire someone to fix it. Ultimately, Y2K caused ordinary people very few problems — but that’s only because there was a huge expenditure of time and resources within the tech business.

The 2038 problem will affect software that uses what’s called a signed 32-bit integer for storing time. The problem arises because 32-bit software can only measure a maximum value of 2,147,483,647 seconds. This is the biggest number you can represent using a 32-bit system.

time is represented as a signed 32-bit integerWhen a bunch of engineers developed the first UNIX computer operating system in the 1970s, they arbitrarily decided that time would be represented as a signed 32-bit integer (or number), and be measured as the number of milliseconds since 12:00:00 a.m. on January 1, 1970.

Glitch says it’s 1970 again

On January 19, 2038 — 2,147,483,647 seconds after January 1, 1970 — these computer programs will exceed the maximum value of time expressible by a 32-bit system using a base 2 binary counting system, and any software that hasn’t been fixed will then wrap back around to zero, thinking that it’s 1970 again.

UNIX time coding has since been incorporated widely into any software or hardware system that needs to measure time.

BI spoke with Jonathan Smith, a Computer and Information Science professor at the University of Pennsylvania for confirmation. The professor confirmed the Year 2038 is a real problem that will affect a specific subset of software that counts on a clock progressing positively. He elaborated:

Most UNIX-based systems use a 32-bit clock that starts at the arbitrary date of 1/1/1970, so adding 68 years gives you a risk of overflow at 2038 … Timers could stop working, scheduled reminders might not occur (e.g., calendar appointments), scheduled updates or backups might not occur, billing intervals might not be calculated correctly

The article concludes that we all need just to switch to higher bit values like 64 bits, which will give a higher maximum. In the last few years, more personal computers have made this shift, especially companies that have already needed to project time past 2038, like banks that need to deal with 30-year mortgages.

64 bitsApple (AAPL) claims that the iPhone 5S is the first 64-bit smartphone. But the 2038 problem applies to both hardware and software, so even if the 5S uses 64 bits, an alarm clock app on the phone needs to be updated as well. (If it’s using a 32-bit system in 2038 it will wake you up in 1970, so to speak.) So the issue is more of a logistical problem than a technical one.

HowStuffWorks reports that some platforms have different dooms-days.

  • IBM (IBM) PC hardware suffers from the Year 2116 problem. For a PC the beginning of time starts at January 1, 1980, and increments by seconds in an unsigned 32-bit integer in a way like UNIX time. By 2116, the integer overflows.
  • Hardware and softwareMicrosoft (MSFT) Windows NT uses a 64-bit integer to track time. However, it uses 100 nanoseconds as its increment and the beginning of time is January 1, 1601, so NT suffers from the Year 2184 problem.
  • On this page, Apple states that the Mac is okay out to the year 29,940!

rb-

The tech industry’s response to Y2K suggests that they will mostly ignore the 2038 issue until the very last minute when it becomes to ignore.  Another example of the pace of global software updates is that a majority of ATM cash machines were still running Windows XP, and thus vulnerable to hackers even though Microsoft discontinued the product in 2007.

Dont worryFortunately, the 2038 problem is somewhat easier to fix than the Y2K problem. Well-written programs can simply be recompiled with a new version of the C-library that uses 8-byte values for the storage format. This is possible because the C-library encapsulates the whole time activity with its own time types and functions (unlike most mainframe programs, which did not standardize their date formats or calculations). So the Year 2038 problem should not be nearly as hard to fix as the Y2K problem was.

Related articles

 

Ralph Bach has been in IT long enough to know better and has blogged from his Bach Seat about IT, careers, and anything else that catches his attention since 2005. You can follow him on LinkedInFacebook, and Twitter. Email the Bach Seat here.