Transcript
Commercial Ethernet services Gwenn AMICE, François MARCOTTE Senior Applications Engineer
CWDM: the Urban technology!
© 2010 EXFO Inc. All rights reserved.!
2
CWDM Topography • Very popular for these types of users: • Enterprise • Cell backhaul • Deep fiber • Single or fiber pair: transmit and receive • Each customer (enterprise or tower) receives a wavelength via an optical add/ drop multiplexer (OADM)
© 2010 EXFO Inc. All rights reserved.!
3
CWDM Technology • Requirements: • • •
Limited network capacity Limited distance (dispersion and power) Low-cost alternative for quick and tolerant bandwidth increase
• • ITU-T G.694.2 • 1261 nm to 1621 nm • 20 nm spacing
© 2010 EXFO Inc. All rights reserved.!
4
CWDM Specifica:ons • Up to 18 channels, typically 16 • 1370 and 1390 nm not used because of water peak • Typical output power: +5 dBm
Tx (λ2)
Rx (λ2)
CWDM ring
Tx (λ3) Tx (λ4) Tx (λ5) Tx (λ6)
40-80 km section
Demultiplexer
• No amplifier
Rx (λ1)
Multiplexer
• Minimum power at receiver: -25 dBm
Tx (λ1)
Rx (λ3) Rx (λ4) Rx (λ5) Rx (λ6)
Tx (λ7)
Rx (λ7)
Tx (λ8)
Rx (λ8)
© 2010 EXFO Inc. All rights reserved.!
5
CWDM Fiber characterization F.A.Q Do I need special tools for fiber characterization for CWDM? No, fiber is fiber, we recommend standard characterization like you would do with any fiber using: Ø IOLM or OTDR for Loss and connections qualification Ø Fiber inspection probe for connector cleanliness Ø Mid-range Dispersion tester.
© 2010 EXFO Inc. All rights reserved.!
6
Testing requirements Do I need to test the Loss of fiber for every single CWDM wavelength? No, fiber characteristic is well known and the IOLM/OTDR test are mainly required for distance, stress, bends, bad splices and connectors qualification. If the fiber is good @1310/1550/1625nm it will also be good @ every 16 CWDM wavelengths!
© 2010 EXFO Inc. All rights reserved.!
7
CWDM Fiber characterization F.A.Q I am worrying about the water peak….Do I need to run a fiber attenuation profile on fiber for CWDM? No, system vendors design their CWDM system in order to deal with legacy fibers, for this reason 1370 and 1390 nm are not deployed. If you truly believe having a trouble with the water peak we suggest you add 1383nm wavelength to your OTDR.
© 2010 EXFO Inc. All rights reserved.!
8
Water peak detection using OTDR @1383nm • OTRD will locate and quantify the Water peak absortion as well as the overall attenuation with high accuracy with no need for calibration or reference • A spectral measurement will only indicate the overall attenuation: bad reference will false the measurement
CWDM Specifications – Chromatic dispersion limits a LOT lower than for DWDM – CWDM is for now limited to 2.5Gbps – Reason: Direct modulation causes CHIRP – CD id important to test, even for 2.5Gbps – CHIRP is induced by the modulation, the higher the bit rate the wider the signal
mw
Ghz
CWDM Fiber characterization F.A.Q Do I need to test Chromatic dispersion on fiber for CWDM? No: if you are not planning anything faster than 1Gbps (1GigE) Yes: If you are planning 10GigE
t z,t
© 2010 EXFO Inc. All rights reserved.!
11
CWDM Fiber characterization F.A.Q Do I need to test Polarization Mode Dispersion on fiber for CWDM? At 10Gbps definitively yes, PMD is a problem on fiber that could impact the transmission at 10Gbps. However PMD testing at 1Gbps is overkill but if you plan to upgrade the GigE service to 10GigE services PMD testing is required
REMEMBER: PMD is a problem!
t z,t
© 2010 EXFO Inc. All rights reserved.!
12
Testing dispersion through CWDM system
© 2010 EXFO Inc. All rights reserved.!
13
Sensitivity
Sensitivity
PM vs. CWDM analyser vs. OSA
Wavelength Wavelength A power meter will mesure the TOTAL POWER
A FOT5200 will mesure the POWER per channel
An OSA will mesure the POWER versus wavelenght Including OSNR & λ
CWDM system activation F.A.Q What kind of source should I use for emulating the different channels? The best CWDM source would be the CWDM SFPs plugged into a CWDM system or an Ethernet analyzer.
© 2010 EXFO Inc. All rights reserved.!
15
CWDM system activation F.A.Q CWDM analyzer OCA or OSA? Which product is the best fit for me? If you are in the process of deploying CWDM technology only, we recommend the FOT-5200 CWDM analyzer. If you are planning to deploy both CWDM and DWDM, the OSA is the way to go!
© 2010 EXFO Inc. All rights reserved.!
16
New Hybrid Network Analyser: OCA Ø Upgradable to a full blown OSA
© 2010 EXFO Inc. All rights reserved.!
17
Troubleshooting CWDM
Head-end
§ § § §
Each customer has a dedicated wavelength Each customer has different distance from head-end Conventional OTDR cannot test through CWDM system Need CWDM OTDR for system installation, maintenance and trouble shooting
CWDM OTDRs are now IOLM ready? CWDM OTDR/IOLM signals will follow the right path and reach each premises, based on its respective wavelength
CWDM evolution…. Ø Migrating from 1GigE to 10GigE – Dispersion tolerances more stringent Ø Mixing CWDM and DWDM
Ref: CISCO & Aurora
© 2010 EXFO Inc. All rights reserved.!
20
Planet plugged In
In 60 secondes
RFC-2544 Objectives: Maximum performances without any errors q q q q
Throughput - Bandwidth Back-to-back Frame Loss Latency
Frame distributions q 64B,128B, 256B, 512B, 1024B, 1518B q Time T
100% 80% 60% 40%
T= 60sec
time
ITU-T Y.1564 Phase 1: Service Configuration Test Phase 1: Service Configuration Test Objective: Validate the network configuration of each defined service (rate limiting, traffic shaping and QoS).
Methodology: For each service, a ramp test is used to gradually reach and exceed the CIR. All KPIs are measured against a threshold.
Ramp repeated for each service CIR
CIR + EIR
All SLA parameters are measured at each step (FD, FDV, frame loss (OOS)) with pass/fail results. ~ 1 minute per service © 2010 EXFO Inc. All rights reserved.!
24
ITU-T Y.1564 Phase 2: Service Performance Test Phase 2: Service Performance Test Objective: Validate the quality of service of each defined service and prove SLA conformance.
Methodology: All services are generated at once to their CIRs, and all KPIs are measured for all services.
5 Mbit/s
10 Mbit/s
All pass/fail results of the SLA parameters are measured throughout the test (i.e., throughput, frame delay, frame loss, frame delay variation and OOS)
2.5 Mbit/s
From 30 s (can be extended for long-term tests) © 2010 EXFO Inc. All rights reserved.!
25
Up to 10 frame size can be test Frame 64B
CIR + EIR
Frame 128B
CIR + EIR
Frame 1518B
CIR + EIR
© 2010 EXFO Inc. All rights reserved.!
26
RFC 2544 or SAM Y.1564 Testing point Transparent network
RFC 2544: 2 hours – more ?? Y.1564: 3 min
Test Scenario
Advanced Ethernet testing
Service - QoS
Troubleshooting
TCP
TEST
SAM Y.1564
SAM + TCP test
UDP
SAM or Frame analyser
© 2010 EXFO Inc. All rights reserved.!
28
Testing Ethernet over CWDM Test point
EtherBERT: 24 hours
Testing Ethernet over CWDM (2) 10G
Test point
BV-3100 Up to 16 Tests
Switch 1G 1G 1G
1G
Testing Ethernet over CWDM
Loopback vs. Bidirectional Measurement (1)
Direction
Customer
Network Maintenance Appliance
Carrier Ethernet Network
RX Rate
Portable in loopback device
Test Rate
70 Mbit/s
100 Mbit/s
70 Mbit/s
100 Mbit/s
100 Mbit/s
30 Mbit/s
100 Mbit/s
30 Mbit/s
Test Rate Network Rate Policing
Network RX Rate Rate Policing
Measurement from maintenance appliance to maintenance appliance = 30 Mbit/s Are you sure about the accuracy of your measurement? © 2010 EXFO Inc. All rights reserved.!
32
Loopback vs. Bidirectional Measurement (2)
Direction
Customer
Network Maintenance Appliance
Carrier Ethernet Network
RX Rate
Test Rate
70 Mbit/s
100 Mbit/s
70 Mbit/s
100 Mbit/s
100 Mbit/s
30 Mbit/s
100 Mbit/s
30 Mbit/s
Portable
Test Rate Network Rate Policing
Network RX Rate Rate Policing
Measurement from maintenance appliance to portable = 70 Mbit/s Measurement from portable to maintenance appliance = 30 Mbit/s © 2010 EXFO Inc. All rights reserved.!
33
Performance Guide for Metro Ethernet
Performance Attributes
Real time application Ie: VOIP
High Lower priority data priority data
Pt-Pt
Pt-Pt
Pt-Pt
Latency one way (ms)
≤8
≤ 20
≤ 37
Packet delay variation, Jitter (ms)
≤2
≤8
N/S
Frame Loss Rate (ratio)
≤0.001% i.e. 10-5
≤ 0.001% i.e. 10-5
≤ 0.1% i.e. 10-3
Questions?
© 2010 EXFO Inc. All rights reserved.!
35
