Upload
stevie391
View
193
Download
12
Embed Size (px)
Citation preview
Jitter and Wander Testing inHigh-Speed Telecom NetworksAndreas Alpert
ITSF 2006
AgendaNetwork evolutionJitter and wander aspectsApplication scenariosCase studyConclusion
© 2006 JDSU. All rights reserved. 3
Evolution of optical transport data rates
Source: Infonetics Research, February 2006
Increase of data rate by 4 and by 10
CY82 CY86 CY92 CY97
155M
622M
2.5G
10G
40G
?
2.7G
10.7G
43G
Hardware Ports
SDH/SONET
OTN G.709
10BASE
100BASE
1GE
10GE
100GE Ethernet
40GE ?
40G
© 2006 JDSU. All rights reserved. 4
Combining TDM and packet switched networks
SDH/SONET
CWDM DWDM DWDMOTN 43G
NG SDH/SONET
FC
Ethernet
GigE
OTN 10.7G
10GE
10GE
High-speed service 10GE:WAN (9.9G), WAN OTN (10.7G), LAN (10.3G), LAN OTN (11.1G)
© 2006 JDSU. All rights reserved. 5
Transport of IP over optical medium
3 play
IP
Ethernet
GFP
NG SDH/SONET
Optical medium (WDM, dark fiber)
VC/LCAS
PPP
Services
CC
Routing
Transport
Physical
OTN
© 2006 JDSU. All rights reserved. 6
SDH network synchronization – well implemented
GPS controlled PRCSSU (TNC or LNC)SEC (G.811)
OTN is not required to transport synchronization
OTN allows the transport of synchronization via SDH client connections
© 2006 JDSU. All rights reserved. 7
Timing distribution via Ethernet – an example
Timing is distributed from PRC to IWF across the packet switched network
Ethernet switches are part of the synchronization network
TDM TDMPacket SwitchedNetwork
Synchronizationnetwork
G.8261/Annex A:
PRC
IWF IWF
Interworkingfunction
AgendaNetwork evolutionJitter and wander aspectsApplication scenariosCase studyConclusion
© 2006 JDSU. All rights reserved. 9
Difference between jitter and wander
Frequency
10 Hz
Measurement time
Time
+ peak
- peak
MTIE
Observation timeTime
TIEat tend
TIE max
TIE min
Jitterin UI
peakpeak
Wanderin ns
Wander range Jitter range
© 2006 JDSU. All rights reserved. 10
ITU-T Recommendations for OTN/SDH/PDH
Primary Reference Clocks (PRC)
G.812
G.813
G.823
G.824
G.811
Synchronisation Supply Unit (SSU)
SDH Equipment Slave Clocks (SEC)
PDH Network Interfaces (2 Mb/s)
PDH Network Interfaces (1.5 Mb/s)
G.825 SDH Network Interfaces
Jitter/wander generation, tolerance and transfer
G.8251 OTN Network Interfaces
Definitions and terminologyG.810
G.783 SDH Equipment Functional Blocks
© 2006 JDSU. All rights reserved. 11
Rec. for synchronization over packet networks
Network time protocol (NTP) – accuracy in ms
IETF RFC2030
IETF RFC1305
Simple network time protocol (SNTP) – accuracy in s
Timing and synchronization in packet networksITU-T G.8261
Precision Clock Synchronization ProtocolIEEE 1588
Precision Time Protocol (PTP) is able to synchronize distributed clockswith an accuracy of less than one microsecond
poor precision
TDM over packets needs to be compliant to existing TDM timing standards
physical layer (layer 1)vs.
in-band (layer 2)
How to transport the timing information
ITU
IEEE
© 2006 JDSU. All rights reserved. 12
ITU-T Recommendations for test equipment
Jitter and wander measuring equipment for SDH
O.173
O.172
Jitter measuring equipment for OTN
Jitter and wander measuring equipment for PDHO.171
MOD
CLK
PTN IF
ext.sync.
Generator
IF CR
REF PD
ext.ref.
Analyzer (jitter, wander)
FIL
HP+LPLP
PPRMSTIEMTIETDEV
AgendaNetwork evolutionJitter and wander aspectsApplication scenariosCase studyConclusion
© 2006 JDSU. All rights reserved. 14
Jitter and wander applications
Jitter/wander SDH/SONET/PDHOTN (wander not required)
Ethernet
Generation BERT scan (Bathtub curve)
Tolerance Stressed eye
Transfer not applicable
DUT
Network Network
© 2006 JDSU. All rights reserved. 15
Requirements for wander test equipment
Generation of wander amplitudes/frequenciesAnalysis of TIE, MTIE, TDEV, frequency offset and drift rateBuilt-in masks that comply with ITU-T, Telcordia, ETSI and ANSISample rates 1/ s, 30/ s, 60/ s, 1000/ sReference input for clock 1.5/ 2/ 5/ 10MHz and data 1.5/ 2Mbs
Referenceclock
Output port:e.g. STM-N
Networkor NE
© 2006 JDSU. All rights reserved. 16
Commissioning test configurations
Tests on network element (NE) level prior to installationTests on synchronisation interfaces after installation
NE
Timing reference
STM-N
Wander generation
NE
Timing reference
STM-N
Wander tolerance
NE
Timing reference
Wander transfer
Synch. IF
© 2006 JDSU. All rights reserved. 17
Acceptance test configurations
Tests on interfaces handing over timing between telecom networksTests on reference clock sources
REF
Timing reference
REF
Network
Timing reference
Characterizationof timing transport
Characterizationof timing quality
Network
Synchronized configuration
STM-N STM-N
Non-synchronized configuration
© 2006 JDSU. All rights reserved. 18
Test configuration for asynchronous signals
AsyncClock
Timing reference
Characterizationof network wanderwith elimination ofclock offset ±x ppm
±x ppm
TIE with frequency offsetOffset corrected TIE
Network
Maximum relative time interval error (MRTIE)
SignalSource
e.g. 2Mb/s
© 2006 JDSU. All rights reserved. 19
Wander on OTN network elements
OTN NE
Mapper (byte stuffing)
Optical TransportNetwork
OTN (G.709)
OTN 10.7G
SDH 10G
De- mapper
FIFO
OTN NE
10G IF
10G Clock
SDH 10G
Filling level
OTN 10.7G
Mapping processWander on OTU clockleads to byte stuffingwhich introduces wanderinto recovered 10G
Long term OTUk bit rate shiftleads to adjustments of 10Gdepending on the filling level ofthe FIFO used for bit rate adaption
The OTN physical layer is not required to transport network synchronization There is no need for specification of network wander limits (ITU-T G.8251)Our recommendation: Wander measurement at OTU interfaces to ensure proper operation at SDH interfaces
AgendaNetwork evolutionJitter and wander aspectsApplication scenariosCase studyConclusion
© 2006 JDSU. All rights reserved. 21
Examples for wander analysis
Whitenoise
Frequencyoffset
TIE MTIE/TDEVMODSRC
© 2006 JDSU. All rights reserved. 22
Examples for wander analysis
Sine0.1 Hz
Sine+ noise+ offset
TIE MTIE/TDEVMODSRC
AgendaNetwork evolutionJitter and wander aspectsApplication scenariosCase studyConclusion
© 2006 JDSU. All rights reserved. 24
Conclusion
Carriers try to converge new and legacyservices on to a single platform, a packet switched network such as metro Ethernet. Thiswill allow them to reduce both capital and operating costs.Synchronization is essential to enable real-time and high-speed transmission for telecom networks.As telecom networks migrate to packet networks, the debate is about how the„synchronization” is passed from one networknode to another.Several standard bodies are working on that topic.Carriers need to evaluate and decide whether“in-band” or “physical layer” is the preferredtechnique to transport the timing information.
© 2006 JDSU. All rights reserved. 25
Thank youThank youfor your attentionfor your attention
Q & A