Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Amazon Time Sync Service

aws.amazon.com

Post reply on HN
Print view View original

71–80 of 94 posts

Re: Amazon Time Sync Service

#71

Post by froh »

NavinF wrote: post #52
deelowe wrote: post #49

Earlier quoted context omitted.

Amazon time sync likely also requires extremely accurate clocks on the system board as well to prevent there from being too much drift.

You don’t even need a tcxo. If you don’t care about holdover, connecting the pps pin from a $5 GPS module to an $15 SBC will give you a time source that is significantly better than anything you can serve over NTP to your clients. NTP is always gonna be the weak link for a service like this.

That's when you move from NTP to PTP, which in a LAN provides sub microsecond accuracy.

https://en.m.wikipedia.org/wiki/Precision_Time_Protocol

With that underneath you need just one gps clock reciever for the whole LAN, and PTP capable Ethernet adapter and switches throughout, of course. Which most are, these days.

Re: Amazon Time Sync Service

#72

Post by froh »

deelowe wrote: post #58
NavinF wrote: post #52

Earlier quoted context omitted.

You don’t even need a tcxo. If you don’t care about holdover, connecting the pps pin from a $5 GPS module to an $15 SBC will give you a time source that is significantly better than anything you can serve over NTP to your clients. NTP is always gonna be the weak link for a service like this.

I was referring to the amazon time sync service which I imagine has pretty tight timing requirements (similar to spanner), not NTP. Also, how exactly would GPS work at scale for 10,000s of servers within 1000s of racks within a DC?

By using PTP in the data center LAN to distribute the GPS time locally.

https://en.m.wikipedia.org/wiki/Precision_Time_Protocol

Re: Amazon Time Sync Service

#73

Post by mrep »

dekhn wrote: post #27
skj wrote: post #11

Earlier quoted context omitted.

Google uses map reduce extensively... where it's appropriate. True time helps with things like spanner transactions. It's just a totally different use case.

The tech lead of the Google MapReduce team (which no longer exists) just received their award for turning down mapreduce. IIRC it was officially done 5 years ago. However I believe the code to delete MR was never checked in and I'm not sure if there are still users. MapReduce was used at Google for highly inappropriate things. For example, the machine learning system I worked on, Sibyl https://www.datanami.com/2014/0…

MapReduce was deprecated because flume [0] the successor is better but it does practically the same thing and flume is used massively. I believe dataflow is the public google cloud version.

[0]: https://research.google/pubs/pub35650/

Re: Amazon Time Sync Service

#74

Post by TecoAndJix »

oblio wrote: post #69
pugworthy wrote: post #62

Earlier quoted context omitted.

Nah. They just figure out eventually one of them is kind of wonky and ignore it. Or maybe both are, so they end up just looking at the stove clock. Also you can't tell a story just with the Aesop ending - you have to tell the fable and END with that line.

I can't find the fable.

I found “Segals Law” https://en.m.wikipedia.org/wiki/Segal%27s_law

Re: Amazon Time Sync Service

#75

Post by oblio »

TecoAndJix wrote: post #74
oblio wrote: post #69

Earlier quoted context omitted.

I can't find the fable.

I found “Segals Law” https://en.m.wikipedia.org/wiki/Segal%27s_law

Ah, no, the misunderstanding was with something else. Apparently "Aesop ending" is a keyword/crossword definition. It means "the moral of the story" :-)

Re: Amazon Time Sync Service

#76

Post by ergl »

londons_explore wrote: post #6

It's really hard to use these API's correctly. Remember... your CPU can halt at any time for any number of milliseconds. That means simple things like: upperBound, lowerBound = readTime() if (upperBound Are incorrect... There is no guarantee that the 'if' statement didn't take many milliseconds, and that the stuff didn't end up happening after the deadline. It's also very easy to write code that works, but is theoret…

To be fair your example would be unsafe with and kind of time API.

The APIs provided by TrueTime and Time Sync are useful to compare two events, each with their own uncertainty intervals. Then you can be sure if any event "happened before" the other, or if they're concurrent.

Re: Amazon Time Sync Service

#77

Post by NavinF »

froh wrote: post #71
NavinF wrote: post #52

Earlier quoted context omitted.

You don’t even need a tcxo. If you don’t care about holdover, connecting the pps pin from a $5 GPS module to an $15 SBC will give you a time source that is significantly better than anything you can serve over NTP to your clients. NTP is always gonna be the weak link for a service like this.

That's when you move from NTP to PTP, which in a LAN provides sub microsecond accuracy. https://en.m.wikipedia.org/wiki/Precision_Time_Protocol With that underneath you need just one gps clock reciever for the whole LAN, and PTP capable Ethernet adapter and switches throughout, of course. Which most are, these days.

Do you have experience with this? Playing around with PTP has been on my TODO list for quite a while, but it seems to be quite a hassle.

At home, I already have a GPS module with its PPS pin connected to an Pi running an NTP server. In my data center, I use clock.sjc.he.net and that's about as good as NTP gets. Certainly good enough for one-way latency measurements.

Are there any SBCs with hardware timestamping and a PPS input? I've heard mixed reviews of the BeagleBone Black as a PTP server. Then again, even if I get that working I have no way to run PTP in the data center since there's no GPS reception in the cabinets.

Re: Amazon Time Sync Service

#78

Post by bigsparky »

foota wrote: post #66
strimp099 wrote: post #34

Can someone please explain to me what this is all about like I’m five years old?

The combination of a time and guaranteed error bound can be used as a primitive in distributed systems. You can get timestamps from requests coming from different systems and relate them in a meaningful way. At least, in a false negative fashion. E.g., if A occurs at T + 100ms with 1ms uncertainty and B occurs at T + 105ms with 2ms uncertainty you can conclude that A happened before B. This does not allow you to alwa…

You must know some pretty smart five year olds!

Re: Amazon Time Sync Service

#79

Post by Faaak »

babelfish wrote: post #2

Using atomic clocks too, nice. From an older post: > It uses a fleet of redundant satellite-connected and atomic clocks in each Region to deliver time derived from these highly accurate reference clocks. https://aws.amazon.com/blogs/mt/manage-amazon-ec2-instance-c...

so... GPS ?

Re: Amazon Time Sync Service

#80

Post by deelowe »

NavinF wrote: post #61
deelowe wrote: post #58

Earlier quoted context omitted.

I was referring to the amazon time sync service which I imagine has pretty tight timing requirements (similar to spanner), not NTP. Also, how exactly would GPS work at scale for 10,000s of servers within 1000s of racks within a DC?

As I understand it, “amazon time sync service” is just a bunch of NTP servers run by Amazon. Their ClockBound library calculates error bounds the same way you would if you were using a different NTP server: |Local Offset| + Root Dispersion + (Root Delay / 2) >how exactly would GPS work at scale for 10,000s of servers GPS module (+ GPSDO + Rb clock optional) -> SBC running NTP server -> machines running NTP client I’m…

I see. This different than what I thought it was. If it is indeed a simple NTP server, this makes sense.
Post reply on HN