As 5G commercial deployment approaches, studies on 5G standardization, technologies, services, ecosystem, and deployment modes are being deepened. It has been proven that services rather than technologies are the main driving forces of network development. Therefore, the functions and performance required for the implementation of 5G networks need to be analyzed from the perspective of application scenarios and service requirements.
eMBB is a deterministic requirement during early 5G deployment, and is also the core to drive the accelerated growth of the entire 5G industry. All early standard formulation, commercial use, testing, and verification of products are based on such requirement.
This paper analyzes and describes the challenges that operators may face in early 5G deployment, especially the problems encountered in 5G base station (gNB) deployment, the preparation of 5G bearer networks, O&M complexity resulting from network virtualization and slicing, and the impact on operators’ organizational architecture. It also proposes the solutions.
Massive MIMO Featured gNB
In eMBB scenarios, the core requirement of 5G networks is for a significant increase in access rates. As rich spectrum resources are available on millimeter-wave bands (26 GHz, 28 GHz or above), it is easy to achieve a cell access rate of over 10 Gbps. However, network coverage and construction cost are big constraints. Considering the industry chain and 5G candidate spectrum around the world, sub-6G bands have gradually become the preferred spectrum resource of mobile operators in their early 5G deployment, and 3.5 GHz is their top-priority choice. According to Ovum, by the second quarter of 2017, there are 15 markets around the globe that have allocated 3.5 GHz band to 5G or have used 3.5 GHz band for 5G testing and verification.
In the fully competitive market, it is difficult for each operator to obtain more than 100 MHz from the 3.5 GHz spectrum. Therefore, utilizing the limited spectrum resources to deliver expected 5G access rates is the greatest challenge for operators. Moreover, as the early 5G deployment focuses on building macrocell sites, operators are still highly concerned about site density and construction costs. Using the existing 4G infrastructure to achieve decent coverage at the 3.5 GHz band and reduce the number of 5G gNB sites is also a major concern of operators.
Massive MIMO is the most important core 5G technology to improve spectrum efficiency (Fig. 1). With the precise beam forming capability provided by large-scale antenna arrays, Massive MIMO allows multiple users to share the same spectrum resource through space division multiplexing, thus increasing cell throughput several times. In the 5G field test in Guangzhou, ZTE and China Mobile demonstrated over 6 Gbps downlink cell peak throughput by utilizing the MU-MIMO feature of Massive MIMO (using the 100 MHz bandwidth of the 3.5 GHz band). Massive MIMO can significantly improve not only system capacity but also network coverage. This is of great importance in supporting continuous coverage of 5G networks and reducing site density and construction costs.
It is commonly agreed that 3.5 GHz 5G NR provides much poorer coverage than 2.6 GHz LTE due to many factors such as propagation loss, penetration loss, and time-division multiplexing issue. To have the same coverage, more base stations need to be constructed. ZTE has demonstrated through theoretical simulations and field tests that the 3.5 GHz 5G network can be co-sited with the 2.6 GHz LTE network based on Massive MIMO and dual-antenna CPEs. This will help operators greatly reduce investments in their early 5G deployment.
In addition to hardware design, the Massive MIMO deployment also has requirements on its algorithm and performance optimization. With the large-scale commercial deployment of Pre5G Massive MIMO worldwide, ZTE has gained rich experience in channel estimation and multi-user multi-stream algorithm optimization. This has laid a solid foundation for ZTE to reap the first-mover advantages in 5G. Currently, ZTE’s 5G products at low frequency bands have distinct advantages in technical maturity and performance. Compared with the traditional deployment mode (antennas + base stations), Massive MIMO base stations will undergo remarkable changes in installation, debugging, and optimization. The large-scale commercial use of Pre5G Massive MIMO will make ZTE well prepared for the whole 5G deployment process.
Massive MIMO can not only increase several times the spectral efficiency of a 5G network but also enhance coverage capabilities of a 3.5 GHz 5G network. It also supports co-sited 4G networks to help operators reduce initial 5G deployment costs and difficulty.
FlexE-Based 5G Transport
Transport is the first concern for 5G deployment. With the accelerated process of 5G standards and commercial 5G systems and terminals, time is running short for transport network reconstruction and upgrade. Operators will invest a lot in reconstructing existing transport networks before deploying 5G networks. Mobile operators must select the most proper reconstruction plan for their transport networks by fully taking into account the 5G network architecture and gNB deployment modes.
Since 5G uses broader spectrum and Massive MIMO, the CPRI interfaces are no longer applicable and need to be re-split. Considering some baseband functions are centralized and virtualized, the baseband part is also divided into distributed units (DUs) and centralized units (CUs). Their deployment modes are quite flexible, depending on service requirements and transport networks. DUs can be deployed with AAUs on the site side, converged to the convergence node to form a resource pool, or deployed with CUs in the central office (Fig. 2).
Compared with 4G networks, 5G eMBB can achieve over 100-fold increase in access capability, which raises higher requirements on transport bandwidth. Due to the flexible deployment requirements of 5G CUs and DUs, transport networks have to deal with fronthaul, midhaul, and backhaul scenarios. In the future, 5G will support massive IoT access, ultra-reliable low-latency communication, and inter-DC on-demand connectivity after network virtualization is completed. It can be anticipated that 5G transport networks will face the challenges of unprecedented complexity and flexibility.
Currently, the standardization of 5G transport is being advanced. ZTE has proposed its FlexE-based 5G transport solution that can flexibly handle hybrid access scenarios involving fronthaul, midhaul, and backhaul and easily deal with inter-DC on-demand connectivity after network virtualization is completed. The solution can also deliver extra-high broadband access capability to eMBB services through flexible slices, and ultra-low-latency forwarding channels to latency-sensitive services. FlexE has therefore become an ideal 5G-oriented transport solution.
Cloud Native-Based 5GC
Virtualization is an inevitable trend of future core networks, while end-to-end network slicing allows operators to easily cope with differentiated requirements in vertical industries and provide customized services for users in different fields in an agile way. After network virtualization, the ratio of software in place of hardware will gradually increase. The operation mode of multi-network coexistence and multi-service concurrency will rely more on intelligent algorithms, and the network O&M team shall gradually transform into the lifecycle management team focusing on new service development and network slicing.
ZTE’s 5G core network (5GC) uses cloud native and micro service architecture to meet different needs of high bandwidth, massive MTC, low latency and high reliability as required by different services in the same network. End-to-end network slicing allows for flexible function orchestration, on-demand resource scheduling and diverse service capabilities among the core, transport and access networks. Moreover, ZTE’s carrier-grade DevOps system enables operators to develop, test, operate and optimize new services. This greatly shortens time to market for new services and reduces O&M complexity.
The standardization of 5GC is later than the 5G RAN schedule. Some operators may plan to first deploy 5G NR and then introduce 5GC at the second phase. Other operators may consider building a complete 5G network at the beginning. In either mode, the virtualization of core networks does not rely on the progress of 5GC standardization. After core networks are virtualized, both EPC and 5GC will be loaded onto the virtualization platform in the form of software applications. Therefore, 5GC standards can be smoothly supported through software upgrade after they are frozen. The virtualization of core networks at the 4G phase can not only reduce O&M costs of 4G networks, but also be well prepared to introduce 5G networks.
As the process of 5G commercial deployment speeds up, most operators are actively involved in verifying new technologies, developing future-oriented evolution strategies, and exploring innovative ICT business modes. In facing the challenges of 5G deployment at this stage, it is of vital importance to select an appropriate path at a right pace.
5G network deployment will cover access, transport and core networks as well as the network management system. ZTE has taken the lead in pre-commercial 5G end-to-end solutions and system performance, carrying out the world’s largest 5G pre-commercial field test. The company has also worked with China Mobile and Qualcomm to complete the world’s first R15-compliant end-to-end interoperability test. These advantages will help global leading operators become the first to launch their 5G commercial services and seize the market opportunities in the 5G era.
5G commercial deployment, Massive MIMO, FlexE, Cloud Native, gNB