[ntp:questions] Re: Question about RFC-2030 SNTP Version4

Christopher J. Holland msnews at microsoft.com
Sun Aug 3 12:45:05 UTC 2003


Hi,

>It sounds as if some other software is doing that for you. You
> seem to be working backwards.  I would assume that you need the whole 64
> bits for an accurate calculation.
Cyberkit is giving me the Round Trip time.
I have to add the Round Trip time to the Originate Time to get a 64 Bit
Number.

> What actual time?  There are four separate events in sending and receiving
> a packet.  Which one are you trying to measure?
There are only 2 packets. A Transmit and a Receive. (UDP)
There is a network delay that the formula takes into account.

I guess I understand that Picture now. Bizarre. I didn't want to get into
the nitty gritty of NTP.

d = (T4 - T1) - (T3 - T2)  = Difference in the 2 trip times
t = (1/2) ((T2-T1)+(T3-T4)) = Average of the 2 trip times. (C->S) and (S->C)

Now that we know d and t, what to we do to calculate the Actual time?
Actual Time = Reference Clock + d + t ???

I think that it could be done this way....
The Time spent at the Server is T3 -T2.
1. TA =  (1/2)(T3 - T2)
2. TB = Server to Client (T4 - T3)
Actual Time = Reference Clock + TA + TB

I don't understand what Root Delay is.
Root Delay: This is a 32-bit signed fixed-point number indicating the total
roundtrip delay to the primary reference source
Cyberkit is showing 0 for this number.
I am thinking this value is T3-T2, but that is not equal to 0.
I'm not terribly sure what Root Delay refers to.

Root Dispersion is the Error.
Answer = plus or minus Root Dispersion

Thanks,
"David J Taylor" <david-taylor at blueyonder.co.uk> wrote in message
news:xK3Xa.1240$l_2.13934629 at news-text.cableinet.net...
> > Then T4 is the time at which the packet is received from the Server.
>
> T4 is the formula, yes.  It is not in the packet.
>
> > Actually, I noticed that Cyberkit keeps track of the Packet Times.
> > Round Trip is calculated by the Client  (00:00:00:13:429 -
> 00:00:00:14:639)
> > = 201 mS.
> >
> > Since, I know the Originating TimeStamp, I can just add the Round Trip
> to
> > get T4?
> > T4 = T1 + 201mS
> > T4 = 14039630149786221412 + 201000
> > T4 = ???
> > (Not sure if I should use the Fractional part, or the whole 64 bits in
> my
> > calculations?)
>
> If you are handling the packet back from the server, you would note T4
> when the packet arrived back with you, and use it to compute the round
> trip.  It sounds as if some other software is doing that for you.  You
> seem to be working backwards.  I would assume that you need the whole 64
> bits for an accurate calculation.
>
> > Once I am able to calculate d and t,
> > Does the actual Time = T1(Originate Time) + Reference Clock Error +
> Local
> > Clock Offset?
>
> What actual time?  There are four separate events in sending and receiving
> a packet.  Which one are you trying to measure?
>
> Cheers,
> David
>
>





More information about the questions mailing list