In article <[EMAIL PROTECTED]> [EMAIL PROTECTED] wrote: > question: "What are the advantages/disadvantages in going with a > 'completely proprietary' product such as Domain Time II, versus the open > source NTPD?"
I imagine the normal advantages of a proprietary solution: - someone to blame; - someone to put things right, but only if you pay them enough to waive the disclaimers of warranty; - a user interface that makes it easier to do simple things; - a user interface that can be shown off to non-technical people. However, in spite of their saying that "Domain Time II" is an open protocol, they fail to provide a link to the specification and I couldn't find it on a quick search. Without that information, there is no hope of making a valid judgement. They also don't implement NTP, only SNTP, with what seems to be less sophisticated clock disciplining, but more sophisticated than some SNTP implementations. Their typical figures for "NTP" are actually outside the step threshold for a full NTP implementation. It's almost as though they were doing a FUD on NTP. (Typical for NTP on loaded Windows is under 30ms, maybe under 5ms for an unloaded system.) > The level of accuracy I need would be the same as that required for, > say, video stream packets (definitely below 20ms). Stop. Wrong platform. This isn't achievable on a loaded Windows system. Even if it were, you would need to write special time keeping code in the application that read the TSC to interpolate between ticks. Basically, if you have to know the performance and are pushing the limits for the platform, you need either published source code (not necessarily "open source") or a contractual commitment to a specific performance level. Could you please give threading software a chance by actually replying to the article rather than starting a new one and copying the old article into it. _______________________________________________ questions mailing list [email protected] https://lists.ntp.isc.org/mailman/listinfo/questions
