In the next five years, 5G network (the fifth generation mobile communication technology) will become a hot spot in the field of information and communication. The construction and upgrading of bearer network will increase the demand for optical devices in telecom network. At the same time, in the era of cloud computing, the demand for data center network has increased explosively, and Internet enterprises have entered the period of IDC(Internet Data Center) demand expansion. Therefore, the demand for high-speed optical modules such as 25G transceivers and 100G transceivers has increased rapidly.
1. Characteristics of 5G network
5G mobile network provides three major types of services: enhanced mobile broadband (eMBB), ultra-reliable low latency communication (uRLLC), and massive machine-type communication (mMTC). The performance of different services varies greatly: eMBB service for traditional mobile communication, with large bandwidth; uRLLC service for industrial automation and other real-time control applications, with low latency and high reliability; mMTC for IoT mMTC is for IoT applications, with many connections and small traffic.
The 5G radio access network (RAN) is divided into active antenna unit (AAU), distribution unit (DU) and centralized unit (CU), and the core network gradually changes from centralized deployment in 3G/4G era to cloud-based and distributed deployment, with different service core networks sinking to different locations to meet the requirements of low latency services and improve user experience.
- Large bandwidth
The base station bandwidth depends on the wireless spectrum bandwidth, spectrum efficiency, number of antennas and other parameters configuration, 64 TR 100M bandwidth base station. For a base station with 64 TR 100M bandwidth, the peak bandwidth can reach 6Gbit/s and the average bandwidth can reach 3Gbit/s. According to the definition of the International Telecommunication Union (ITU), the maximum peak bandwidth of 5G base station can reach 20Gbit/s. However, in practice, the base station rate is difficult to reach the maximum peak rate. In addition, considering cost, power and other factors, 5G base stations will coexist in many types, and the base station bandwidth will exist from 1Gbit/s to 20Gbit/s.
5G base stations are divided into high-frequency base stations and low-frequency base stations: at the initial stage, 5G low-frequency base stations and 4G base stations are deployed together; In the mature stage, the density of 5G low-frequency base stations is equivalent to that of 4G base stations, while 5G high-frequency base stations are mainly used for heat supplement. In the initial stage, the scale is small, but some require 25Ge interface access.
- Low latency
The latency of different 5G services varies greatly. The 3rd Generation Partnership Project (3GPP) TR 38.913 defines eMBB end-to-end (E2E) latency as 10ms and uRLLC as 1ms, with eMBB having a null delay of 4ms and uRLLC having a null delay of 0.5ms; however, for different uRLLC services, 3GPP TS 22.261 V16.0.0 gives different delay definitions for different uRLLC services.
- Traffic mesh
The CU (Centralized Unit) and DU (Distribution Unit)of 5G are deployed flexibly and can be separated or combined. The separated CU and DU have the characteristics of many-to-one and one-to-many, and there are dual homing and redundancy requirements. According to the step-by-step introduction of eMBB, uRLLC and mMTC services, the core network is gradually transferred from centralized deployment to distributed deployment. There is a many-to-many relationship between the CU and core network, and there is traffic interaction between core networks. The meshing trend of service traffic is obvious in 5G era.
- Network slicing
The next generation mobile networks (NGMN), IMT 2020 and the third generation partner (3GPP) have all proposed the network slicing architecture of 5G network based on Software Defined Network (SDN) and Network Function Virtualization (NFV). The network slicing can provide the basis for network innovation and rapid deployment of services in the future. At the same time, network slicing service can provide management isolation, resource isolation, computing isolation, forwarding isolation, control isolation and other characteristic services. The isolation of different resources can be configured flexibly to meet the differentiated requirements of different types of business security, reliability, key performance indicators (KPIs) and ensure business security and service quality.
- Gradual evolution from NSA to SA
5G network construction mode is divided into standalone (SA) mode and non-standalone (NSA) mode. Under SA mode, new wireless and core 5G networks, 4G network and 5G network operate independently. NSA is a gradually evolving network technology scheme. It is enhanced only on the 4G network through the existing resources of 4G, provides services for 5G in the way of local capacity expansion, and gradually evolves to 5G with the continuous maturity of 5G services.
The 5G bearer network uses 10Ge and 25Ge interfaces to access the base station at the user network interface (UNI). Moreover, the 25Ge, 50Ge and 100Ge interconnection technologies will be introduced at the network side, and the metropolitan area may evolve to 200Ge and 400Ge links in the future. Therefore, the 5G network construction will need fiber optical transceivers with the characteristics of integration, miniaturization, high rate, long distance, low cost and low power consumption, such as 25G optical modules and 100G optical module.
2. Trends of 25G optical module and 100G optical module
- Introduction to common types of 25G optical modules
In 2020, 5G starts to embark on its commercial journey, and 25G optical modules are now widely used in the construction of 5G signaling base stations. In the 4G era, single-mode 10G optical modules were mainly used in the fronthaul network, while in the 5G era, 25G SFP28 optical modules are widely used in the construction of 5G fronthaul base stations due to their optimal input/output (I/O) performance and fiber capacity, which is 2.5 times higher than that of 10G Ethernet.
- 25G SFP28 SR optical module: center wavelength of 850nm, duplex LC interface, multi-mode, operating temperature from 0°C to 70°C, transmit optical power from -8.4 to +2.4dBm. It is widely used in data center 25G Ethernet and 5G forward CPRI/eCPRI network. The maximum transmission distance of the SFP28 SRmodule is 70m over OM3 fiber and 100m over OM4 fiber.
- 25G SFP28 LRoptical module: center wavelength of 1310nm, duplex LC interface, single-mode, supports commercial grade operating temperature (0°C ~ 70°C) and industrial grade operating temperature (-45°C ~ 85°C), transmits optical power from -8.4 to +2.4dBm. It’s generally paired with single mode patch cable OS2. It is widely used in data center 25G Ethernet and 5G forward transmission CPRI/eCPRI network, and the longest transmission distance of SFP28 LR module is 10KM.
- 25G SFP28 ERoptical module: Center wavelength of 1310nm, duplex LC interface, single-mode, supports commercial grade operating temperature range (0°C~70°C). Generally used with single-mode patch cable OS2. It is widely used in data center 25G Ethernet and 5G forward CPRI/eCPRI networks with the longest transmission distance up to 40KM.
- 25G SFP28 BIDIoptical module: center wavelength of 1270nm-Tx, 1330nm-Rx, simplex LC interface, single-mode, supports industrial grade operating temperature range (-45°C to 85°C). Generally used with single-mode patch cable OS2. It is widely used in data center 25G Ethernet and 5G forward CPRI/eCPRI networks, with the longest transmission distance up to 40KM.
- 25G SFP28 WDMoptical module: A) 25G SFP28 DWDM optical module: Its fixed wavelength is C18-C61 (interval 100GHz), duplex LC interface, single mode, supports industrial grade operating temperature range (-20°C to 85°C). It is generally used with single-mode patch cable OS2. It is widely used in data center 25G Ethernet and 5G forward CPRI/eCPRI networks, with the longest transmission distance reaching 10KM. B) 25G SFP28 CWDMoptical module: optional wavelength 1470nm-1570nm (interval 20nm), duplex LC interface, single mode, supports industrial grade operating temperature range (-20°C~85°C). It is generally used with single-mode patch cable OS2. It is widely used in data center 25G Ethernet and 5G forward CPRI/eCPRI networks with the longest transmission distance up to 10KM.
- Introduction to common types of 100G optical modules
- 100G QSFP28 Optical Transceiver：Itis the mainstream 100G optical modulein the market today, and the package form is QSFP28 (Quad Small Form Factor Pluggable). Because of its advantages of small size, high port density and low power consumption, it is loved by the majority of manufacturers. The transmission wavelength of 100G QSFP28 optical module is 1310nm, and its transmission distance can reach 2km with a single-mode 12 core MTP (MPO) connector. The size of QSFP28 transceiver is smaller than other 100G modules.
- 100G CFP Optical ModuleSeries(CFP/CFP2/CFP4)：The CFP MSA is the first industry standard to support 40G and 100G Ethernet optical terminals. The size of CFP is the largestamong all the 100G CFP optical modules, the CFP2 is one-half the size of the CFP, and the CFP4 is one-quarter the size of the CFP. It should be reminded that CFP/CFP2/CFP4 optical modules cannot be used interchangeably, but they can be used simultaneously in the same system.
3. The development of 25G/100G optical modules
The rate of optical module has always been the focus of market attention. At present, the factors promoting the rate upgrading of optical module mainly come from several aspects. On the one hand, to achieve 100 megabit fiber to the home, from the access layer to enhance the pressure of the optical interface, driving the demand for high-speed rate optical modules. On the other hand, with the deployment of 5G, operators need to deploy larger bandwidth to achieve high-traffic data applications, such as telemedicine, VR, 4K video, etc. Therefore, mobile networks must have higher rates at all levels, which also drives the development of 25G optical modules and 100G optical modules. In addition, besides the huge demand for optical modules in carrier networks, the accelerated construction of cloud computing data centers has also raised the demand for 25G optical modules and 100G optical modules. From the market perspective, the demand for 100G optical modules brought by the new construction and transformation of global data center networks will scale up and the high-speed optical module market will remain high boom.