Cloudified Edge DC Promotes Network Reconstruction and Telecom Service Transformation

Telecom operators are facing the challenges and opportunities of transforming from traditional networks to new networks such as SDN/NFV. With the promotion of virtualization technologies, DC will become the core carrier of future telecom networks. Because different telecom services have different features and coverage ranges, unified DCs cannot meet requirements. The services deployed in DC must be classified into different levels to meet the cost and service requirements. Based on the distributed and multi-level DC architecture, the SDN/NFV network is gradually becoming a consensus of network evolution among major operators.

Edge DC becomes the focus concerned by operators

Edge DC mainly covers the local network layer, meets the requirements of local service access, shortens service response time, focuses on low latency and improves user experience. With the evolution of 5G, it becomes a trend to deploy computing capabilities to edge nodes. Along with the rapid use of local content and real-time information of edge nodes, network services, services, and applications can be distributed and downloaded more quickly, effectively reducing the pressure in mobile core networks, improving the telecom service environment, and allowing users to enjoy better network experience.

In the 5G era, a large number of services will be deployed in edge DC. A large number of local edge DCs that are close to users play an important role in improving operators' competitiveness. They are the core advantages and important assets of telecom operators. The construction of edge DCs has naturally become the focus concerned by operators. 

Features of edge DC

Edge services have the features of low latency, high bandwidth, and localization. Therefore, edge DC that carries these services has the following features:

• The VM performance demand is high 

Edge DC is close to users. To ensure user experience, the services deployed in edge DC require high-performance VMs/containers for implementing services.

• The O&M workload is huge

In accordance with the coverage range and service deployment features, there are a large number of edge DCs, which inevitably brings a lot of O&M work.

• Physical conditions can be limited

An operator has a large number of local traditional communication equipment rooms. Considering cost and reuse, a large number of edge DCs will be deployed in these traditional equipment rooms. These equipment rooms are far behind DCs in terms of key indicators such as area, load-bearing capacity, power supply capacity, and energy efficiency, so they need reconstruction. The deployment of edge DCs has physical limitations because some equipment rooms are relatively poor in conditions and difficult in reconstruction.

Edge DC solution

To address the pain points of edge DC, ZTE has launched the edge DC solution (see Figure 1). Ultra-integration, hardware acceleration and other technologies can improve the reliability and performance of VMs, and unified centralized O&M will greatly reduce O&M workload. The use of dedicated integrated hardware breaks through physical conditions, making the deployment of edge DCs more convenient.

Figure 1 Edge DC Solution Architecture

High integration and hardware acceleration of compute/storage/network 

ZTE Edge DC Solution introduces the ultra-integration technology to integrate computing and storage on the same physical platform. On the basis of ensuring the storage reliability, it improves the storage access performance and meets the high performance requirements of edge DCs for I/O. This mode is based on the existing hardware modules. It can implement fast expansion and compatibility between different modules, and the expansion process is very simple. The virtualization and ultra-integration infrastructure provides efficient, concise, reliable, and low-cost solutions for the reconstruction of edge DCs.

Services are diversified on edge DC. The processing performance of general hardware for specific service requirements is low, and even cannot meet service requirements. Therefore, dedicated hardware acceleration needs to be considered. Some services need intensive calculation, such as 5G CU PDCP air interface encryption/decryption and MEC positioning algorithm; some services require traffic forwarding, such as 5G UPF/xGW and local MEC traffic distribution; some services are related to high-performance video technologies such as MEC AR/VR and video live broadcast.

ZTE supports a complete hardware acceleration solution to achieve higher server resource utilization and meet the requirements of edge DC. For example, ZTE launches the intelligent network card solution with low latency, high forwarding capability and high cost performance. It unloads the data plane function of the software vSwitch, which needs a large number of physical server CPUs, to the hardware of the intelligent network card. The released CPU resources after being unloaded can be used to deploy the service applications.

Centralized O&M and local O&M management system

The O&M management system uses a two-level management system: centralized O&M management and local management. It makes full use of centralized management to perform remote operations, reducing maintenance workload and costs. In addition, it provides local O&M to facilitate maintenance of single sites to meet customers' different requirements.

As ZTE's cloud management platform for multiple DCs, TECS Director aims at simplifying the management and operation of cloud DCs. TECS Director integrates resource management, monitoring and O&M through connection deployment in the homogeneous or heterogeneous cloud environments of different DCs. In addition, the centralized backup and restoration system, version management system, and patch system can implement data backup and restoration, version update, and patch operations for all edge DCs in an area. This improves the reliability of DCs and reduces construction costs.

Due to the service diversity of edge DCs and different application scenarios, the scale and requirements of edge DCs are different. Therefore, edge DCs should be deployed in accordance with local conditions. ZTE provides solutions such as distributed OpenStack deployment (deployment in remote access sites), standard OpenStack/PaaS deployment, simplified OpenStack/PaaS platform deployment and virtualized management system, so that customers can select appropriate platforms as needed.

Dedicated integrated hardware

When some edge DCs are affected by space, power, air conditioning, and load bearing, and cannot be deployed due to the limitations of general hardware, ZTE provides dedicated integrated hardware to match the telecom equipment room conditions in the existing network, thus minimizing the need for equipment room reconstruction. The integrated hardware balances the size, weight, power consumption, performance, and scalability, and matches the constraints of the edge computing environment, providing carrier-class reliability.

With the emergence of new services such as edge computing and fog computing, the reconstruction and transformation of various traditional services under the 5G evolution makes the edge DC that bears these services a more popular topic. ZTE Edge DC Solution meets the deployment scenarios of various edge DCs, makes cloud construction better serve services, promotes telecom industry transformation and network reconstruction, and helps operators embrace the new 5G era.