100G Ethernet has prevailed in the data center with the advancement of tech and the development of networks. Naturally, various 100G applications spring up with it, in which the relatively hot application is 100G optical transceiver module. Although most of people has had a good knowledge of 100G optics, and if exploring it from other different perspectives, it will be found that there are still many details that we don’t know. Then in the next content, something different will be introduced by Gigalight.
1. The Background of 100G Optical Transceivers
The earliest one type of 100G optical transceiver emerged in the form factor of CFP in 2010. At that time, IEEE released three standards(SR10, LR4 and ER4) for 100G optical modules, respectively focusing on the 100ms, 10kms and 40kms transmission. Then the IEEE standard added new 100G SR4 project, but in 2013 it did not reach consensus. In 2016, the 100G optical transceiver modules for various data centers mostly adopted the 25Gbps Serdes program. After that, the 100G optical modules that use the 50Gbps Serdes plan slowly appeared.
2. The Development Progress of 100G Optics: From CFP to QSFP28
At present market, the 100G optical transceiver module include CXP, CFP, CFP2, CFP4, CPAK and QSFP28. Among them, QSFP28 demonstrates its great superiority and will lead to denser optics and further price reductions. The QSFP28 increases front-panel density by 250% over QSFP+. The increase in panel density is even more dramatic when compared with some of the other 100G transceiver module: 450% versus the CFP2 and 360% versus the CPAK. In addition, the surge of QSFP28 shipments will be one of the factors to change the market from 40G to 100G, according to the report of IHS. QSFP28 is becoming the universal form factor in the data center.
3. The Challenges of 100G Optics
At the process of development of 100G optic fiber transceiver, there exist a series of challenges as follows:
- Channel Distance: The DWDM system supporting the 50GHz wavelength distance is very extensive. The 100G optical module needs to be accorded with the condition of supporting the 50GHz wavelength distance. Therefore, the pattern of high spectral power should be used.
- OSNR (Optical Signal-to-Noise Ratio): Under the same pattern, 100G optical module requires 10dB higher than 10G optical module and 4dB higher than 40G optical modules. Therefore, a low OSNR tolerance code and FEC algorithm with high coding gain are adopted.
- CD margin: Under the same conditions, the dispersion tolerance for 100G optical modules is 1/100 of 10G optical modules, accounting for 16/100 of 40G optical module. It can use dispersion compensation technology, by compensation in the electric field or the optical domain, to complete the dispersion compensation for each wavelength.
- PMD tolerance: Under the same conditions, the PMD (polarization mode dispersion) tolerance of 100G optical modules is 1/10 of that of 10G optical module, accounting for 4/10 of 40G optical module, so one needs to choose coherent reception and digital signal processing.
- Non-linear effects: the non-linear effects of 100G optical module are than 10G / 40G optical module nonlinear effects more complex.
4. The Application Situation and Development Trends
1). Application Situation: 100G Optical Transceiver Module Is More Widely Used
Previously, 100G was primarily installed on high-end core routers and now more are being installed on lower-price edge routers, which significantly reduces the price of 100G optical transceiver. In 2016, the construction of global data center market will keep growing which means that the 100G optics will be applied in a larger scale. Geographically, North America, Europe and Asia-Pacific (especially China) are the main market of 100G transceiver with their increasing demand for deployment of 100G equipment.
2). Development Trends: 100G Optics More Cost-Effective Than Before
The cost of transceiver modules which keep adding over time is one of the main considerations for the whole projects. In other words, the cost of the devices and components may influence the enthusiasm of network upgrade. But, in 2016, the 100G transceivers will be more affordable. On the one hand, the low-cost 100G silicon reaches production and the technology becomes mature. On the other hand, the widespread utilization of the 100G devices and a great deal of increases in the Internet traffic are the core to change in the communication infrastructure markets. This reduction in pricing will lead to that 100GE is sold at a price per bit below that of 10GE in the 2018.
In addition to the trend on the cost, there exit these trends. For example, the power consumption tends to be lower than before. The interface port also tend to be higher. The form factor is smaller and so on.
At the progress of 100G optical transceivers from CFP to QSFP28, there exist various challenges in many aspects. Even so, it is believed that these problems will be overcome with the advancement of network tech. It will be more conformed to the markets demands on cost, form factor, power consumption, etc. Let’s expect it together.