5G defines three application scenarios, each requiring a technology generation. Ruitai (Weihai) Electronic Technology Co., Ltd. This technology generation will not be less than 5 years. Assuming that 5G is deployed from 2019, it will take less than 10 years to complete the envisioned deployment of 5G networks. We have to carry on the technical and the application feasibility analysis to this, thus provides the basis for the superstructure to rethink.
5G has three application scenarios
●5G conventional applications of EMBB, defined to achieve the goal of virtual reality technology and ultra-high definition video sharing, anywhere cloud access and up to 1G Internet bandwidth.
● Next to 5G are URLLC, which is low latency autonomous driving and industrial Internet applications.
●5G's third application MMTC, namely machine communication, to achieve the Internet of vehicles and intelligent asset management, that is, to make everything automatically connected to become a reality.
Looking at the three applications of 5G planning, and covering every technology area that human technology has just touched, each of which itself will take a long time to break through. As these applications of 5G definition are reshaping social rules, it brings severe challenges to social redefinition and social management. Just to clarify the issues of social management and politics requires a lot of practice and argument.
Let's set aside for a moment three scenarios of possible applications. It is important to have an achievable optical transmission network as the physical basis for all these applications. We have seen the technical plans of the three major carriers for 5G carrier networks, but the reality is that these plans must be reconsidered in terms of cost, technology, and logic for certain applications. Otherwise 5G will be a long way off.
We've seen at least two carriers embed a lot of tunable technology in 5G. But the technology itself has a maturity and application paradox. For example, China Telecom has embedded 25G tunable optical module technology and ROADM technology.
It will take at least two to three years for 25G tunable optical modules to move from the laboratory to the industrial environment. The problem is that fixed wavelengths and fixed lasers are more reasonable and economical to use. Arguably, a tunable deployment is flexible and simple to maintain. We said that the so-called flexible deployment is the problem of tags and tags, all the module wavelengths have been written into the EEprom or MCU, the background can be clearly identified automatically, from the background did not find the meaning of flexible. When it comes to the foreground, i.e. the construction site, there is no difference between the action of inserting all light modules with the same tag and the action of inserting all light modules with clearly marked wavelengths. The physical port of the machine can match any wavelength of light module, that is, the light module itself can be plugged into the machine at random. At the very least, the argument goes, if we send a batch of identical light modules to the construction site and light modules of different wavelengths to the construction site, the latter could be mistaken. The problem is that it weakens the ability of optical module manufacturers. All optical module manufacturers' operation and logistics have been trained for a long time to adhere to the principle of zero-error quality. Ease of maintenance is also non-existent. Maintenance itself is about product performance and reliability. There is little dispute in the industry that fixed-wavelength lasers are not only cost-effective but also low in quality maintenance. On the other hand, the mass adoption of tunable lasers will bring a lot of maintenance and failure. Since the tunable laser is temperature-dependent, the temperature drift of the module itself is logically inconsistent with the requirements of 5G application environment in harsh environments, so there are also serious risks at this level. As for ROADM technology, there is little doubt that the cost, loss, and delay cannot be adopted by the core network (refer to another article of mine "ROADM Technology Shape a Generation of Technological Illness" for reference). Tune technology itself is used for link backup. We don't have to be ignorant to pay for this expensive technology. Dynamically tunable networks and the strict stability required by 5G networks create conflicts.
PAM4 is a modulation method that multiplies bandwidth usage with constant port density. With the 56G NRZ experiencing technical bottlenecks, PAM4 technology is beginning to take off. This mode of modulation has proved to be completely unproblematic. But there is still a big difference between the actual application and the laboratory, indoor environment and outdoor environment. PAM4 technology in practical applications should require an ideal transmission link, but at low speeds. Currently, due to 56G NRZ problems, PAM4 modulation has been used for high-speed transmission, which brings more nonlinear effects than low-speed transmission. Coherent communication is a reliable technology because it does not change the quality of the electrical signal and does phase processing for the optical signal. The basic principle of PAM4 technology is to use denser levels to transmit more information. The signal distortion of denser levels within the module and on the optical link cannot be solved by PAM4 technology. The adoption of PAM4 DSP or PAM4 CDR is key to the success of this technology. Now we generally look, VCSEL short distance using PAM4 technology + analog CDR is completely feasible, long distance using EML technology +PAM4 (DSP) is also a feasible technical path. The third possible path is Silicon Technology +DML + Analog CDR. Based on the above three judgments, we believe that in the 5G high-speed network, if we cannot successfully introduce Silcion technology, 5G prequel will not be suitable for the introduction of PAM4 technology. In the field of return transmission, PAM4 using EML technology is feasible. One question is whether to use fully coherent technology in 5G back-transmission or to partially add PAM4 to long-distance products, which of course depends on cost and link distance.
Here we must discuss an original question, is it multiplexing or optical reuse? The multiplexing end of 5G networks uses higher rates, from 100G to 200G to 400GG, and the same goal can be achieved with spatial WDM. Assuming that the application of 5G does not care about the space of the machine room, the traditional NRZ technology + DWDM passive technology + coherence can achieve the established goal of 5G return transmission. If the deployment density is concerned, the PAM4 technology + coherence is a solution. We need to understand that the delay caused by PAM4 technology is much greater than with NRZ technology. This may also pose a risk for some applications.
Whether single wave 100G technology can be used in 5G optical transmission network is still an unsolved problem. It is generally better to dismiss the application for the time being and leave it to a longer term technical observation.
The three applications defined by 5G are ambiguous. The basic 5G accelerated application is basically feasible and can benefit the basic goal of most people. The second application, which involves autonomous driving, would presumably have to be an AD hoc network, which would be relatively complex if subdivided into a massive 5G network. Huge planning creates delays and redundancy. In addition, the big risk of autonomous driving is the accuracy of local computing and sensing, and the network transmission depends on the density and reliability of the network. The intensity of 5G networks matches this goal, but all this is done using very mature technology rather than introducing experiments. The third application of 5G, the Internet of Things itself, is not so much about speed and does not require huge amounts of bandwidth. This application is a high-speed regional or industry network. So 5G only needs a high-speed interface. One more thing to understand about the Internet of Everything is how necessary it is that when machines and machines become separate kingdoms, humans are actually excluded. The application of this layer is still too early for us to understand the practical implications of this layer.
If they want to carry everything for humanity, 5G networks will have to plan for a lot of local data centers. Operators of course have this condition and foundation. But we must understand that 5G network, as a physical network, cannot have the convenience and low cost of management of the Internet. Some technical workers' vision of 5G goes beyond the reality and achievements of technology application. Cramming in some supposedly advanced technology has nothing to do with achieving 5G goals. Back to the optical network itself, what it wants to achieve is density and speed, and bandwidth allocation. In general, we see the application of 5G network EMBB as a speed war, we need to do the job of maximizing the bandwidth available to the terminal, this network can not be very rigorous. The 5G URLLC network needs the high-speed bandwidth interface of 5G and the flexible allocation of network bandwidth, which basically uses OTN technology. The application of MMTC involves the transformation of everything. The Internet of Everything is more based on the networking of local clouds and public clouds, and its transmission has nothing to do with 5G.
Ruitai (Weihai) Electronics
Company Address: No.3 Building, Central Europe Industrial Park, Jinhua North Road, Lingang Economic and Technological Development Zone, Weihai City, Shandong Province
Statement: please read and understand this statement before browsing this website. Some pictures and words originated from the Internet, if there is infringement, please contact to delete。