104 TimeAnalyzer 7x00 User’s Guide
098-00208-000 Revision A – February, 2010
Chapter 5 Analyzing
Analysis of PDV Measurements
Calculations and Masks
TimeMonitor PDV includes two computations for packet timing data PDV analysis,
packetTDEV and packetZTIE. These computations are based on - but are not
identical to - floor-based calculations such as minTDEV, percentileTDEV, and ZTIE
with minimum transit time packet selection.
Further, checking or unchecking "Optimal" has a direct bearing on the packetTDEV
and packetZTIE calculations. The calculations and the resulting plots are affected
by selecting or unselecting "Optimal". In general, the result is improved by selecting
"Optimal". In all cases the result is the same or better when "Optimal" is selected.
As described above, with "Optimal" unchecked, the plot results are an indication of
how a generic PTP slave will perform. When the Optimal checkbox is selected,
additional analysis is invoked to indicate how a PTP slave with more advanced
packet-timing algorithm will perform under identical network conditions. Such a PTP
client can tolerate more degraded network conditions and still meet the
performance mask.
PacketTDEV
The minTDEV and percentileTDEV calculations, which are the basis for the
packetTDEV calculations, are described in detail in the TimeMonitor Analyzer User
Guide. The definitions are reproduced here, first for minTDEV and then for
percentileTDEV.
The definition of minTDEV is:
where
The percentileTDEV calculation is a generalization of the minTDEV calculation.
Like the minTDEV calculation, it is applicable to packet delay data. While the
minTDEV calculation involves the selection of a single minimum delay "lucky
packet", the percentileTDEV finds a set of minimum delay packets and averages
them together. It is a special case of the bandTDEV calculation. For further
information, refer to the TimeMonitor Analyzer User Guide.
The packetTDEV calculation is used as a predictor for how a reasonably
well-designed PTP slave should perform from the standpoint of TDEV under the
network conditions under study. The PTP slave is evaluated by measuring the
telecom signal, such as a DS1 (1.544Mbit) or E1 (2.048Mbit) signal. The TDEV
performance would then be ascertained by applying the standard TDEV calculation
to the DS1 or E1 measurement data. If the packetTDEV result meets the
packetTDEV sync mask, for example, a PTP slave would be expected to meet the
corresponding telecom sync mask, such as the G.823 sync mask.
2
min
min
min
6
1
min
_
2
2
)
(
min
)
(
i
x
n
i
x
n
i
x
TDEV
x
1
min
min
n
i
j
i
for
x
i
x
j
