CALNEX PARAGON-X

Testing 1588v2 PTP

Introducing Calnex Solutions Ltd

Company founded in January 2006.

Executive team with over 100 years of experience
in telecom test instrumentation.
Rapporteur of the ITU-T study group 15 - Q13
First customer shipments February 2008.
- Tier1 Service Providers
- Tier1 & Tier2 Equipment Manufacturers
- Component & Clocking Vendors
De-facto Test Equipment for 1588v2, Sync-E, CES,
NTP testing & Eth-OAM.

02
Components, Clocks,
Systems Suppliers Service Providers
Labs, etc.

Telecom
Chunghwa
Example Calnex Paragon Customers
1588v2 and SyncE in the Network
CORE EDGE AGGREGATION ACCESS
PRC or PTP
1588v2 GM BC/TC/Slave

PTP BC/TC MPLS Sync-E

1588v2 GM
Cell Site
Slave
Routers
PTP BC/TC
Carrier Sync-E
Ethernet BC/TC/Slave

PTP Slave
ONU
P-OTN or BC/TC GPON
P-OTN or
MPLS or OLT
MPLS or
EoS Carrier Ethernet TC
Sync-E
Sync-E BC/TC
Sync-E
Slave

BC/TC
xDSL, etc.)
G.984.2 Sync-E
PTP
ITU-T standards
Part 1 Testing Network Elements
Ordinary Clocks
(1588v2 Slave Clocks)
A network with no BCs/TCs
PDV Accumulation

Master Router Router Router Router Slave

Clock Clock

In a network with non 1588-aware switches, PDV and Asymmetry can be significant.
Slave clock recovery is a challenge.
Tests have shown Frequency recovery is easier but Phase/Time recovery is a challenge
Test the Slave Clocks robustness to PDV and Asymmetry

Company Confidential 8
Test the Slave Clock

1. Apply PDV
(G.8261 test cases)

2. Measure E1/T1
(to G.8261.1, etc.)

3. Measure 1pps, ToD

(vs. specified limits)

ITU-T Mask
Real-world testing

Capture PDV profiles from . . . then replay the real-world PDV profiles back in your lab.
live or trial networks . . . MKR-1:x=894.874691063,
y=0.000360070
MKR-2:x=894.874691063,
Delta: x=0.000000000,
y=0.000000000
x=895.4775188
35
y=0.000444870
0.000645879
y=0.000360070

Time Interval
Error (TIE)
vs Nominal
(seconds)

Core Base
Network Station -0.000296349
894.478017335 894.977768110 895.477518885
T (a) Offset=0.002 ppm

Master Slave
Clock Clock

010
Boundary Clocks (BCs)
Boundary Clocks

Boundary Clocks reduce PDV

accumulation by:
Terminating the PTP flow and
recovering the reference timing.
Generating a new PTP flow using the
Q1
local time reference, (locked to the
recovered time).
Clock
Q2
There is no direct transfer of PDV
Slave Master
from input to output.
Qn
A Boundary Clock is in effect a back-
to-back Slave+Master.

13 13
A network with 1588v2 BCs
PDV PDV PDV PDV PDV

Master Slave
Clock
BC BC BC BC Clock

Boundary Clocks (BCs) are back-to-back Slave-Masters

BCs recover and re-generate the 1588v2 clocking
With a network of BCs, PDV contribution (per hop) is only from BC and link
PDV experienced by Slave is minimised
Test the BCs Output Noise, Noise Transfer and Noise Tolerance

14
 Why Test Boundary Clocks?

Potential Sources of PDV:

a) Clock Wander
Each BC recovers the clock and re-
Q1
generates a new timing signal. This
can lead to the introduction of low
Q2
frequency clock wander,
Qn
Chains of BCs can lead to the
accumulation of low-frequency clock
wander.
TIE

TIE

b) High-freq. PDV from BC

Time Time
PDV from BC Output Buffer Queue;
Input Noise Output PDV
12 s for 1514 byte packet
525 s for 64kB Jumbo packet
PDV from other internal queues.
Affected by other High Priority Traffic?
Company Confidential 15
Test the Boundary Clock

GPS

ITU-T G.8273.2 will specify the

performance of a BC.
Ref

1588v2 + PDV Noise generation

BC
Noise tolerance
1pps ,ToD, E1/T1/BITS
Noise transfer
1588v2 + PDV
Phase Transient Response

Company Confidential 16
Transparent Clocks (TCs)
Transparent Clocks
Transparent Clocks reduce PDV by;
Calculating the time a PTP packet
resides in the TC device (in nsec) and
insert the value into the
CorrectionField.
Using the CorrectionField, the Slave or
terminating BC can effectively remove
the PDV introduced by the TC.

Q1

Q2

Qn

Packet Delay in TC Device

inserted into correctionField
at output of Transparent Clock device
18 18
 A network with 1588v2 TCs
PDV Accumulation

Master TC Slave
Clock
TC TC TC Clock

PDV is written by each TC into

CorrectionField and this accumulates,
so CorrectionField = PDV Accumulation
at the End Slave

End Slave removes

PDV Accumulation
using CorrectionField

19
Test the TCs Accuracy by measuring CorrectionField error
Testing TC Accuracy
1. Measure the packet-by-packet latency
across the TC.
2. Determine the change to the
correctionField value for each message.
Traffic from Traffic Generator
for Congestion noise testing 3. Accuracy is the difference in the actual
latency compared to the change in
Master Slave
Correctionfield value.
`1`
Clock `1`
Clock
TC
Measure impact of correctionField on
Sync PDV.
1. Vary traffic packet size.
2. Vary traffic priority.
Optical Splitter or 3. Vary traffic utilisation.
Electrical Tap
Repeat for Sync & Del_req PDV.
Test in 1-Step and 2-Step modes.
Accuracy Spec is < 50ns (IEEE C.37.238)
Calnex Paragon-X measurement accuracy
is 5ns
20
TC Accuracy example results

35nsec pk-to-pk per TC

(Network of 2 TCs)

Company Confidential 21
1588v2 Application Notes
@ calnexsol.com

Company Confidential 22
Part 2 Testing Networks
Step 1 - Determine Network SLA

pktMTIE
pktTDEV
MAFE
FFO/FPP
etc

Determine the correct Metric and

limits for each slave
Use these limits for Network SLA
Step 2 Measure SLA in the Network and
Troubleshoot Issues

1588v2
NTP
Sync-E
ITU-T / PDH
CES G.823/G.824 Mask
Network
FAILED
Node B
E1/T1 MTIE/TDEV
GPS
MKR-1:x=894.874691063, y=0.000360070 Delta: x=0.000000000, x=895.477518835
MKR-2:x=894.874691063, y=0.000360070 y=0.000000000 y=0.000444870

0.00064588

Time
Interval
Error (TIE)
vs Nominal
(seconds)
1588v2 PDV

894.478017335 894.977768110 895.477518885

T (a) Offset=0.002ppm FAILED
ToD 1.5s limit

1pps & Time of Day (ToD)

1588v2 Analysis Sync-E G.8261 Mask

FAILED

Sync-E MTIE/TDEV
ALL AT THE
SAME TIME
1588v2/NTP PDV & Metrics
025
 PTP Analysis
Graphical measurements

Company Confidential 27
PDV Distribution
Probability Distribution Function

Company Confidential 28
PDV Distribution
Cumulative Delay Function

Company Confidential 29
 PDV Metrics
Floor Packet Population (G.8261.1)
Packet Delay Variation network limit at
point C of figure 3/G.8261.1 (for HRM-1)
Floor Packet Percent (FPP)
Window interval W = 200s
Fixed cluster range = 150s (from floor delay)
At least 1% of packets must fall in this cluster.
PDV Packet metrics Pass FPP
(Single microwave hop + multiple other hops)

Company Confidential 31
PDV Packet metrics Fail FPP
(Network included 10+ Microwave hops)

Company Confidential 32
E1 analysis

Company Confidential 33
1pps Analysis with Pass/Fail limits

Company Confidential 34
Real-World Results
BC/TC Trial Network Test
Traffic from G.8261 Appendix VI:
Traffic Generator Test Case 13
Traffic Model 2

Sw Sw Sw Sw
Master Slave
#1 #2 #3 #n

PDV E1 1pps
Ethernet Connection
Congestion traffic

~
Example Network Configurations;
1. 9 switches, No On-path Support, No SyncE PTP for Frequency & PTP for Phase
2. 9 switches, all BC mode, SyncE SyncE for Frequency & PTP for Phase
3. 5 switches, all BC mode, No SyncE PTP for Frequency & PTP for Phase
4. 9 switches, all BC mode, No SyncE PTP for Frequency & PTP for Phase
5. 9 switches, all TC mode, SyncE SyncE for Frequency & PTP for Phase
6. 9 switches, all TC mode, No SyncE PTP for Frequency & PTP for Phase 36
Results:
BC networks

Test Set-up PDV at E1 wander 1pps

input to (MTIE @ (pk-to-pk)
slave 5000sec) Sync PDV: 105nsec pk-pk
PEC + 9*Sw 86sec 2.44sec 2.70sec
EEC + 9*BC 0.055sec 0.019sec 0.028sec
PEC + 5*BC 0.070sec 0.188usec 0.176sec
PEC + 9*TC 0.105sec 0.210sec 0.220sec

Observations when BCs utilised;

SyncE (EEC) + PTP gave the best results.
BCs reduce the impact of congestion traffic,
but congestion can still impact the transfer of frequency &/or Phase.
37
Results:
TC networks

Test Set-up PDV at E1 wander 1pps

PDV: 86sec pk-pk, without CF
input to (MTIE @ (pk-to-pk)
slave 5000sec)
PEC + 9*Sw 86sec 2.44sec 2.70sec
EEC + 9*TC 245nsec 0.137sec 0.112sec
PEC + 9*TC (86sec)* 1.10sec 1.75sec

Observations when TCs utilised;

SyncE (EEC) + PTP gave the best results.
PDV: 0.245sec pk-pk, with CF
TCs reduce the impact of congestion traffic,
but congestion can still impact the transfer of
frequency &/or Phase.
38
Why Calnex?
Trust us, Calnex knows Sync
Organisation-wide understanding of Sync and Testing
Leadership at Standards bodies
De-facto 1588v2 and SyncE test equipment its what your customer uses
The ONLY Full 1588v2 and SyncE Conformance tester
G.8261 etc 1588v2 tests
G.8262 SyncE Jitter AND Wander
G.8264 SyncE ESMC tests
Clear Pass/Fail Analysis and Troubleshooting features
Packet Metrics pktMTIE, pktTDEV, MAFE, PDF, CDF, FPP, FFO, etc, etc.
PDV graphs is mutliple formats for clear and thorough analysis
1pps and Time of Day
E1/T1/BITS MTIE/TDEV
Dedicated 1588v2 BC and TC one-box solution
Complete G.8273.2 and other BC tests
The only way to test TC accuracy to nanoseconds
39
THANK YOU
calnexsol.com