A few years ago, one of our interns was curious about system time keeping features in computer systems. This intern was also the proud owner of an inexpensive Radio-Controlled Clock. The intern wondered why computer motherboards were not just equipped with a chip like the ones in the RCC so that their system time would always be correct.
I posted a question about this on the CentOS email list and received more responses than those postings about problems with the new Firefox release. I must have really struck a very sensitive system time nerve.
This large response was a bit of a surprise and included a bunch of time related horror stories. It became clear why using an RCC chip on motherboards would NOT be a good idea. GPS network time servers seemed to be a preferred choice.
All of our bedrooms have Radio-Controlled Clocks. At 5:30 this morning, half of the clocks displayed the correct time. The other half of the clocks were incorrectly showing a time one hour ahead. Maybe this is one more piece of evidence to reject using an RCC time base for computers, at lease in thestate of Arizona.
On Sun, 2020-03-08 at 17:59 +0000, Chris Olson via CentOS wrote:
A few years ago, one of our interns was curious about system time keeping features in computer systems. This intern was also the proud owner of an inexpensive Radio-Controlled Clock. The intern wondered why computer motherboards were not just equipped with a chip like the ones in the RCC so that their system time would always be correct.
This large response was a bit of a surprise and included a bunch of time related horror stories. It became clear why using an RCC chip on motherboards would NOT be a good idea. GPS network time servers seemed to be a preferred choice.
The problem with radio time signals is that they just aren't accurate enough. Your bedroom clock needs to be correct to within a minute or so, but they are generally correct to about +/- 5 seconds. That's just not good enough for system times. There's also a massive problem with signal strength in the UK - the (singular) time transmitter is in the middle of the country in Cumbria and in the south it's virtually impossible getting a signal any further than about 2 feet from a window - not a hope of getting anything in an office building!
GPS times also have problems. They are very accurately wrong! The atomic clocks on the satellites haven't been updated since they were launched, so no leap seconds. There are corrections that can be applied once the time has been received but it depends on a knowledge of leap seconds - I think they are currently about 18 seconds out. But they are accurate to about 10-100ns. You also need a decent antenna to get the high accuracy, which again means that you need to be near a window to see the satellites.
Generally much easier to use NTP!
P.
Once upon a time, Pete Biggs pete@biggs.org.uk said:
There's also a massive problem with signal strength in the UK - the (singular) time transmitter is in the middle of the country in Cumbria and in the south it's virtually impossible getting a signal any further than about 2 feet from a window
- not a hope of getting anything in an office building!
There are different systems around the world (WWVB in the US for example), and I don't think there's a system at all in many countries. Also, putting a receiver inside a computer case would pretty much never work for the low radio frequencies used by these systems, so there'd have to be an external antenna (a lot of effort to go to when you could just use network time sources).
Radio clock accuracy is typically in the 100ms range, so is good enough for most people's computer clock usage.
GPS times also have problems. They are very accurately wrong! The atomic clocks on the satellites haven't been updated since they were launched, so no leap seconds.
That is not a problem - GPS time is defined as being continuous, unlike UTC. However, the GPS signal includes the UTC offset, which is updated when UTC applies a leap second, so you can calculate correct UTC from just the radio signal. I'm not as familiar with the GPS alternatives (Galileo, GLONASS, Beidou, and more), but I believe they'd all be the same (a continuous time base, with offsets specified in the data).
Also, again, GPS signals are weak and require an external antenna.
I do have an external GPS receiver and external antenna hooked up to one system at home, so I have a stratum-1 NTP server (probably accurate to about 1µs).
Basically for most, the "chip inside the box to set the clock" is the network chip. :) If you need clock setting on a disconnected network, you can get a dedicated time server.
On Sun, 8 Mar 2020 17:59:16 +0000 (UTC) Chris Olson via CentOS wrote:
why computer motherboards were not just equipped with a chip like the ones in the RCC so that their system time would always be correct.
Digital cinema servers (the gadgets that feed the movie to the projector and sound systems) run on Linux. The movies are shipped to the theatre in an encrypted form and a key is emailed to the theatre that allows the movie to be played from date-and-time to next-date-and-time. Some keys are valid for one day, most are valid for a week, some are valid for six months.
As you can see, the system time on a cinema server is very important. If a key is valid at 11:00am for a movie that's to be shown at 11:00am (which can happen even though it's not a good practice) then if your system time is five or ten minutes slow your showtime gets screwed up because you can't play the movie yet.
And cinema server clocks are notoriously inaccurate. It's one of those ironies that a twenty-thousand dollar server is built with a fifteen cent clock chip in it. And because the showtimes are an important element of the digital cinema key system, the servers all have a "secure clock" in them that can't be adjusted by more than six minutes per year.
(Yes, they're built that way. Cinema servers are the most locked down awful-to-deal-with things that you can possibly imagine outside of digital cinema projectors which are also locked down with the additional feature of anti-tamper devices, and a lot of "hidden stuff" in both servers and projectors. Digital cinema equipment is the only commercial electronic equipment that I know of that is deliberately designed to not do what it's intended to do, i.e. play a movie.)
After learning the pitfalls of this the hard way, most theatres (including mine) now have a computer dedicated to acting as an ntp server in the projection booth to tell the equipment what time it is.
On 3/8/2020 1:55 PM, Frank Cox wrote:
Digital cinema equipment is the only commercial electronic equipment that I know of that is deliberately designed to not do what it's intended to do
When someone asks why their firewall is blocking such and such, I sometimes respond that firewalls are "specialized routers designed to drop traffic [rather than deliver it]"
--On Sunday, March 08, 2020 6:59 PM +0000 Chris Olson via CentOS centos@centos.org wrote:
All of our bedrooms have Radio-Controlled Clocks. At 5:30 this morning, half of the clocks displayed the correct time. The other half of the clocks were incorrectly showing a time one hour ahead.
Probably a result of political changes to when DST transitions occur. We see a new tzdata package fairly regularly because some politician somewhere in the world decided to change when his country would switch between standard time and DST. (It's not just Linux. All operating systems use this package. So that Windows Patch Tuesday may be JUST for a tzdata update!)
The hardware that suffers the most is devices with fixed firmware that can't be updated. Notably, those cheap "atomic" clocks.
So just wait a few weeks for the older clocks to switch on the day the older laws specified.
On Sun, 8 Mar 2020 at 14:01, Chris Olson via CentOS centos@centos.org wrote:
A few years ago, one of our interns was curious about system time keeping features in computer systems. This intern was also the proud owner of an inexpensive Radio-Controlled Clock. The intern wondered why computer motherboards were not just equipped with a chip like the ones in the RCC so that their system time would always be correct.
One issue is that the radio clock frequencies are based on the country they are in. They are also dependent on different characteristics of those frequencies. In the US, the clock measurement which reaches the entire country is the 60 kHz WWWVB. It hugs the ground so does not miss places like the WWV frequencies of 5,10,15 MHz. However WWWVB needs a very long antennae to be used and it can get masked by anything from a kitchen blender or a lot of computer parts. This means that the antennae need to be away from most equipment. The frequencies that most US radio clocks use are the WWV 5, 10 and 15 Mhz. The problem with these frequencies are many: 1. They are shortwave frequencies which bounce off the ionosphere. You may get areas of the country skipped over all different times of day or year or sun sponts. 2. The frequencies are used by a lot of chips. That means you are going drown out a lot of talk from computers. 3. The data in the signals are different per country but due to the ionosphere skipping you might get a different clock source. 4. There is only one signal per country so you can drift out of time easier for various reasons.
https://en.wikipedia.org/wiki/Radio_clock
https://en.wikipedia.org/wiki/WWVB The radiation pattern of WWVB antennas is designed to present a field strength of at least 100 μV/m over most of the continental United States and Southern Canada during some portion of the day. Although this value is well above the thermal noise floor, man-made noise and local interference from a wide range of electronic equipment can easily mask the signal. Positioning receiving antennas away from electronic equipment helps to reduce the effects of local interference.
I posted a question about this on the CentOS email list and received more responses than those postings about problems with the new Firefox release. I must have really struck a very sensitive system time nerve.
This large response was a bit of a surprise and included a bunch of time related horror stories. It became clear why using an RCC chip on motherboards would NOT be a good idea. GPS network time servers seemed to be a preferred choice.
All of our bedrooms have Radio-Controlled Clocks. At 5:30 this morning, half of the clocks displayed the correct time. The other half of the clocks were incorrectly showing a time one hour ahead. Maybe this is one more piece of evidence to reject using an RCC time base for computers, at lease in thestate of Arizona.
CentOS mailing list CentOS@centos.org https://lists.centos.org/mailman/listinfo/centos
-
Blake Hudson -
Chris Adams -
Chris Olson -
Frank Cox -
Kenneth Porter -
Pete Biggs -
Stephen John Smoogen