The advent of the cloud computing era requires high-capacity and high-density data interconnection and exchange processing. The 400G system can further improve the network capacity and reduce the transmission cost per bit on the basis of 100G system, effectively reducing the pressure of continuous growth of business traffic faced by operators. However, to realize 400G optical communication, we need to break through some technical requirements. At present, 400G flexible bandwidth optical network (FBON) technology, 400G optical fiber technology and 400G optical transceiver have been developed relatively mature. This article will take you through these three new technologies and together we will look forward to the popularity of 400G optical communication technology in the future.
- 400G flexible bandwidth optical network (FBON) technology
At present, there are three main implementation schemes for 400G system: ①Build 400Gb/s system through 4-carrier 100G PDM-QPSK. Its advantages are that 100G technology has been commercially available on a large scale, it’s a mature technology with low cost and long span; ②The 400G system is constructed by dual carrier 200G PDM-QPSK/16QAM. The advantages of this scheme are that the spectral efficiency is increased by more than 165%, the system integration is high, the volume is small and the power consumption is low; ③The 400G system is constructed by single carrier 400G PDM-32QAM. This scheme has the highest spectral efficiency and seems to be the best solution. However, due to the limitation of Shannon’s Law, its technical implementation is difficult and costly. The first two schemes are really widely tested and pilot applied.
Although the traditional optical network has many advantages, it also has many shortcomings, mainly as follows: ①Once the wavelength channel is established, the bandwidth of the channel is basically determined and cannot be flexibly adjusted, so it is difficult to adapt to future services and the changing needs of the network, for example, the 100Gb/s channel sometimes only carries 10Gb/s services, which results in a waste of bandwidth resources. ②The optical devices in the existing 10G/40G/100G systems are designed with a fixed frequency interval of 50GHz according to the I-TU-T G.694.1 protocol, i.e., no matter how large the transmission rate is, they have to occupy the 50GHz frequency interval for transmission, which inevitably leads to the waste of frequency resources.
The rapid changes in network applications have greatly contributed to the growth of business traffic, while the demand for fiber optic transmission capacity has become increasingly urgent. Therefore, while introducing higher-order modulation/polarization multiplexing technology to increase the amount of information carried by a single channel, it is also necessary to consider the whole network solution to optimize the spectrum efficiency and thus improve the resource utilization. Therefore, variable bandwidth optical network (FBON) came into being. The key technologies of FBON include: variable bandwidth ROADM technology, variable bandwidth optical transceiver technology, variable bandwidth electrical layer technology and variable bandwidth control layer technology.
- 400G optical fiber technology
The dispersion tolerance for a 400Gb/s signal is only 0.5ps/nm and the polarization mode dispersion (PMD) tolerance is only 0.25ps, which is 1/16 and 1/4 of that for a 100Gb/s signal, respectively. In addition, the effect of nonlinearity and fiber attenuation exceeds the effect of dispersion tolerance and PMD, and will become the main factor affecting the system performance. How to balance the spectral efficiency, transmission distance and system capacity under the constraints of Shannon’s Limit has become a key issue in the deployment of 400Gb/s technology. Increasing the effective area of optical fiber and reducing the attenuation coefficient of optical fiber are the main means to improve the transmission performance and prolong the transmission distance of the system.
The G.652 fiber used in the current network can no longer meet the needs of the future optical transmission network ultra-high speed, large capacity, ultra-long distance transmission, the industry hopes to introduce new fiber technology to provide support for the implementation of 400Gb/s, 1Tb/s and other ultra-100Gb/s technology. Currently, there are three development directions: low/ultra-low loss fiber, large effective area fiber, and combined large effective area and low/ultra-low loss fiber. When domestic manufacturers choose the technology development direction, they basically choose to combine large effective area and low/ultra-low loss development direction, which is the new fiber—G.654. The test results show that compared with the traditional G.652 fiber, the G.654. E fiber can prolong the optical transmission distance by 70% ~ 100%.
- Status of 400G optical transceiver module
In 2018, the 100G data center Ethernet market is in full swing and is the main battleground for major optical module vendors to compete. With the growing maturity of 100G series products of major manufacturers, the shipment volume is rising. The technical threshold of 100G has been crossed and has become the mainstream and first choice for the strong demand of data centers. It is no longer synonymous with high-end optical modules. With the rapid development of technology, in order to meet the growing bandwidth demand of large-scale data center, we have put our hope on the 400G optical transceivers. 400G optical transceivers are widely considered by the industry as a solution that can effectively reduce bandwidth costs.
Before introducing the 400G optical module, it is necessary to know the 100G optical module. There are three criteria for a successful optical module: small size, low power consumption, and flexibility. As we all know, the SFP/SFP+/QSFP+/QSFP28 form factor is used to implement 1G/10G/40G/100G network transmission respectively. In fact, for 100G network, there are four different types of module: CFP, CFP2, CFP4 and QSFP28. As you can see in the figure below, they are getting smaller and smaller in size and power consumption.
The transmission department in telecommunication network needs pluggable transceiver, which also needs to use some special technologies such as coherent detection to cover long-distance transmission. Data centers, on the other hand, need a small, compact device with low power consumption and low unit cost, as their applications are for short-range transmission only (usually up to 2km). In the original 100G transceiver use cases, only the CFP module was available, however, even for medium and long distance ranges, it was not possible to get the power consumption below 12W. As the technology evolved, smaller size CFP2 and smaller CFP4 optical modules emerged. Until today, Coherent technology for 100G and 200G was only available for CFP and CFP2. Meanwhile, the huge demand for data center capacity from GAFA (Google, Apple, Facebook and Amazon) has pushed the QSFP28 form factor to a variety of short-haul applications such as DAC, AOC, SR4, PSM4 and CWDM4. Today, as the technology matures, most 100G applications can use QSFP28 encapsulated optical modules, except for some cases exceeding 40km, coherent technology needs to be used.
400G is the next step in the development of 100G, so 400G transceivers are becoming increasingly important. Based on market conditions, 400G is preferred for internal connectivity in large data centers and is currently less used in the transport sector of telecom networks. Since 400G bit rates require PAM4 modulation, the coverage becomes increasingly challenging and limited to a few kilometers, longer distances will require the use of coherent detection techniques, amplification techniques, dispersion compensation techniques, etc. Overall, there are 2 specifications of optical modules for 400G networks, one dedicated to intra-data center connectivity (Intra-DC) and the other for long distance transmission.
- Data center Intra-DC 400G transceiver
Intra-DC for 400G is available in two sizes: QSFP56-DD (QSFP-DD for QSFP dual density) and OSFP (Octal SFP). The electrical interface for both specifications is 8-channel 50G PAM4, while the optical interface can be either 8 groups of 50G PAM4 lasers or 4 groups of 100G PAM4 lasers.
The QSFP-DD module is defined by the QSFP-DD MSA, while the OSFP module is defined by the OSFP MSA. The OSFP is slightly wider and longer than the QSFP-DD, and therefore takes up more PCB surface area of the switch. OSFP moduel can arrange 32 ports on each 1U panel of the switch, while QSFP-DD can arrange 36 ports. Therefore, in terms of switch bandwidth capacity, QSFP-DD can provide 4 more ports. Moreover, QSFP-DD has backward compatibility with QSFP and QSFP28. From these two points of view, the future of QSFP-DD transceiver module seems brighter. However, the QSFP-DD form factor has higher requirements for the design of the module, which puts forward higher requirements for the internal device package, module performance, power consumption and yield, which may lead to the increase of module cost.
- Long-range transmission 400G transceiver
Compared to QSFP-DD and OSFP, CFP8, as defined by the CFP MSA, is different.
- Allows for up to 24W of power consumption.
- Has a 16x25G NRZ at the electrical interface instead of the 8x50G PAM4 modulation technology used with QSFP-DD and OSFP.
- Has an MDIO management interface for QSFP-DD and OSFP instead of I2C.
The CFP8 has a large footprint and a maximum power consumption of 24W for long-distance transmission applications. Its initial version is 10km with 16 channels of 25G NRZ electrical layer and can be converted to 8 channels of 50G PAM4. With the development of coherent detection technology, a version called CFP8 ZR (80km) will be launched in the future, which opens the door for CFP8 800G.
- Summary
From the technical circuit diagram of IEEE, the 400G optical transmission will come soon. From automated industrial production to vehicle networking, from enterprise networks to carriers, 400G optical communication already points to two application scenarios: one is cloud computing center, and the other is high-speed transmission link. To realize these application scenarios, mature technical support is required. 400G optical transmission system is basically the same as 10Gg system in interface technology, FEC technology and optical amplification technology. Therefore, to successfully transition from 100G to 400G, more breakthroughs need to be made in bandwidth improvement, optical fiber line layout and optical module design. Although affected by the COVID-19 epidemic, technology research and development in various countries has been hindered to some extent, it is believed that with the deep development of 5G technology, the 400G optical communication is still the general trend and is expected to be popularized and applied in the near future.