Deploying networks using CFP transceivers presents a set of challenges that did not exist with previous generations of Ethernet.
Any CFP based 100G Ethernet deployment will commonly use a WDM interface – the only exception to this is in the case of using SR10 CFPs – and this in itself is enough to require a new set of skills and tools to support network deployments.
Typically in an Ethernet turn-up activity one of the most basic tests that would be carried out would be measuring the optical power. This would be measured both on the output of the transceivers being used but also on the received fibre to ensure that the power level is within the appropriate limits for the transceiver.
This simple test is no longer possible when deploying and testing 100G Ethernet circuits.
Due to the use of WDM within the CFP taking a simple power measurement does not prove the correct operation of the CFP.
While it may be possible to measure the aggregate optical power – in terms of verifying that each wavelength is transmitting at the correct power there is no value in doing this. Consider the case of an LR10 CFP where 10 WDM wavelengths are in use. If one of the lasers within the CFP had an output power of only 50% of the expected power how would this present in an aggregate power measurement? It would only represent a 5% change in aggregate power level and this would be unlikely to be recognised by an engineer in the field.
Instead of making these broadband power measurements the early 100G network deployments have relied on the measurement capabilities of the CFP itself. While this approach has allowed technician to turn up and test 100G Ethernet circuits it has several problems.
- Power measurement is not a mandatory function of the CFP. While some CFPs may support it others may not so it is not guaranteed that the measurement is available.
- The CFP is not a calibrated measurement device. Therefore the measurements it makes should be considered to be indicative only.
- The accuracy of the measurement will vary by each product and the technician in the field or the installation manager is unlikely to know the measurement accuracy and judge if it is within acceptable levels.
One of the biggest challenges to 100G Ethernet deployment has been cost and the CFP transceivers represent a significant portion of the cost of the network. As such CFP vendors have been under significant pressure to reduce the price of the CFP and in response to this some vendors are now developing and offering “cost-reduced” CFPs where as much cost as possible is driven out of the device. One of the functions that can be removed from the CFP without affecting the core functionality of the CFP is the power measurement functions.
As this trend continues the availability of the measurement functions in the field will decrease.
So how can these challenges be overcome?
The answer is the use of an instrument that is typically not used in Ethernet turn up activities; the Optical Spectrum Analyser (OSA).
An OSA is typically used for testing DWDM networks, making power, wavelength and optical signal to noise ratio (OSNR) measurements and it is the power and wavelength functions that would be useful in field deployment of CFPs. As the majority of deployed CFPs today are using WDM technology and DWDM CFPs are looming on the horizon an OSA should soon become part of the essential toolkit for field engineers.
During the deployment of a CFP into the network an OSA should be used to verify that the CFP is operating within it’s specifications by measuring the optical power for each wavelength and that the wavelengths themselves are correct. Typically an OSA may be a complex instrument to use and would include many features ad measurements that would not be required during a 100G Ethernet or 100G OTN turn-up activity. This may put the use of the OSA outside of the skill set of the engineer that would typically be turning up Ethernet services. As only the power and wavelength measurement would be required it would be possible that low-end OSA instruments could be used that would minimise the complexity of using the instrument in the field but still additional training may be required for the engineers responsible for deploying the 100G CFP interfaces.
The deployment of WDM CFPs as a network interface therefore present additional turn-up costs compared to traditional interfaces, firstly in terms of the OSA that needs to be used to verify the CFP in place of a low cost power meter and secondly the training and support needed by field engineers in order to carry out OSA measurements on the CFP interface.