nanog mailing list archives
Re: NTP clock source
From: todd glassey <tglassey () earthlink net>
Date: Mon, 29 Mar 2010 09:43:25 -0700
On 3/25/2010 10:40 AM, Dan Shoop wrote:
On Mar 25, 2010, at 8:51 AM, Kyle Bader wrote:Can anyone recommend a solid clock source (stratum 0) that's not overly expensive? The only stuff I've found so far is ESE, can anyone recommend them or conversely has anyone had any problems with their hardware?
Why would you want a S-0 Clock??? You are not a time-space lab, and so you would want something like a stratum-2 time source which comes from a provable provider. There are laws by the way on what are official and non-official sources of time. In the US for instance these are 15 USC 271 and 272, and the right to deploy the time is codified in 15 USC 260 so which source of time is used is important. By the way - if this is a commercial use, try applying the same set of controls to the time provider you are forced to apply to the rest of the outsourcing service providers you rely on...
-- KyleWe have several symmetricom time servers that we use in several location and I'd recommend them very highly. '
Depends on whether you are trying to generate evidence of something or not. If you need to be able to prove the time data is correct, then you will need a process to do that, a process which most all of the time server systems to date choke on pretty badly. Since the GPS-L1 system without SAASM encryption was formally banned by the US Military for use in their systems by an order of the Joint Chiefs of Staff in 1998, its pretty clear that relying on GPS-L1 for anything requiring a sense of digital trust is out. NTP Peer Stats files have virtually no way of associating content to events meaning that the time-tokens which are passed around are lost in the wind, meaning that there is no enduring evidence generated by NTP. The NTP loopstat and peerstat files have just enough info to be dangerous and not enough to prove anything to the existing rules of evidence, so the real issue is whether you have to prove something as a response to a regulatory requirements or you just want to synchronize the systems in question for... Todd
-d ------------------------------------------------------------------------ Dan Shoop Computer Scientist shoop () iwiring net GoogleVoice: 1-646-402-5293 aim: iWiring twitter: @colonelmode
Attachment:
tglassey.vcf
Description:
Current thread:
- Re: NTP clock source, (continued)
- Re: NTP clock source Henry Linneweh (Mar 25)
- Re: NTP clock source Gerry Boudreaux (Mar 25)
- RE: NTP clock source Wallace Keith (Mar 25)
- Re: NTP clock source Dave Hart (Mar 25)
- Re: NTP clock source todd glassey (Mar 26)
- Re: NTP clock source Jon Meek (Mar 25)
- Re: NTP clock source Marty Anstey (Mar 25)
- Re: NTP clock source Lamar Owen (Mar 26)
- RE: NTP clock source Erik Soosalu (Mar 25)
- Re: NTP clock source Dan Shoop (Mar 25)
- Re: NTP clock source todd glassey (Mar 29)
- Re: NTP clock source Valdis . Kletnieks (Mar 29)
- Re: NTP clock source todd glassey (Mar 29)
- Re: NTP clock source Nate Itkin (Mar 25)