Cloud UniCore: Innovative Virtualized Core Network Solution

In the mobile internet era, operators are facing serious challenges from OTTs. ZTE has proposed an M-ICT strategy that is user-centric and future-oriented. The strategy involves in-depth integration of IT, CT, and internet and assists operators to transform their services. Key components of the M-ICT strategy are network function virtualization (NFV) and software-defined network (SDN), both of which have become the focus of industry. With the introduction of NFV, operators are most concerned about:

●    reducing network capex and opex and improving resource utilization while optimizing network performance

●    ensuring IT technology in the virtualized network has carrier-class reliability

●    ensuring rapid, convenient NE deployment, fault locating, and routine O&M in layers provided by multiple vendors

To address the above issues, ZTE proposes a cost-effective, easy-to-manage carrier-class Cloud UniCore virtualized core network solution. This solution involves all NEs in the 2G and 3G core networks and the EPC, IMS, or UDC.

High Price-Performance Ratio

iSDN Gateway: Creating Highly Efficient User-Plane Channels

With x86 servers, pure NFV gateways have strong computational capability and can process data on the control plane. However, compared with dedicated data forwarding devices, they are inferior in forwarding data over the media plane. Gateways such as SGWs and PGWs are the key in the 4G core network, and data forwarding performance is crucial. ZTE proposes the iSDN architecture to achieve an SDN-based gateway. This architecture separates complicated data processing from basic data processing in the media plane, and 80% of user data is forwarded at a wire speed through a high-speed switch. For the remaining data that requires complicated processing, service chain technology is used to schedule and process resources as needed. This not only improves the media-plane performance but also reduces device overhead. In addition, forwarding nodes are distributed by iSDN gateways. The control plane of the gateway selects the optimal forwarding node and path for user traffic according to user locations, service flow characteristics, and user properties (Fig. 1).

Integrated vCN: Simplified Network with High Integration

At present, the common virtualized core network solution still remains in the network element virtualization (NEV) stage. However, ZTE’s integrated vCN solution can freely combine the functional components of various NEs in the core network. This solution can be applied to different scenarios and motivate the network evolution. Through the internal exchange between components and common modules combination, the virtual machine (VM) resource consumption, processing delay, and interworking difficulty between NEs can be reduced. Moreover, the high degree of integration in the integrated vCN solution makes it quicker and more convenient to deploy small networks.

Carrier-Class Reliability

Multi-Layer Reliability Guarantee: Second-Level Fault Recovery

The hypervisor layer is added to the virtualized network. The reliability of a single VM is lower than that of traditional telecom devices, and simple virtualized solutions cannot meet the carrier-class reliability requirements. Therefore, ZTE takes the following measures to improve the reliability of virtualized networks to 99.9999%.

●    Multi-level dispersion and redundancy: POOL geographic redundancy and 1+1 or N+K backup of software and hardware components.

●    Multi-level monitoring and fault self-healing: in-band and out-band monitoring of software and hardware and VMs; insulate and restore faulty resources. By doing so, the system can be rapidly recovered and the service capability can be ensured.

●    Service continuity guarantee: through hot patching, ISSU, VM hot migration, and synchronization, service continuity is guaranteed during the period of fault, O&M, and scaling process.

Intelligent Overload Control to Resist against Network Storm Attacks

A virtualized network combines overload control and scalability to admit new traffic by increasing resources, greatly enhancing the anti-load shocking capability of the network.

●    Load closed-loop monitoring: monitors the CPU load, bandwidth usage, and session number in real time, and adaptively controls the load.

●    A combination of load control and scalability: when the system load does not exceed the license limit, increase processing resources rapidly to admit new traffic. Furthermore, the control plane and media plane can be independently stretched to adapt to different overload scenarios.

●    Overload control: when the system load reaches the license limit, if an overload event occurs, overload control is initiated to suppress the ratio of new transactions. Meanwhile, overload information can be sent to adjacent NEs to ensure that services can still be effectively processed.

Ease of Maintenance

Big-Data Closed-Loop Troubleshooting Solutions

The NFVO, VNFM, VIM, and virtualized layer are newly added to a virtualized network, so the complicated structure makes it more difficult to locate faults. ZTE’s big-data closed-loop troubleshooting solutions locate faults rapidly and accuratly, and the fault locating period is shortened from 4 hours to 15 minutes.

●    Accurate: monitoring hardware, software, and the virtualization of these from a multi-level perspective, and thus data can be collected on fault spots and faulty processes across the whole system. This makes data more complete and enables alarms to be more accurately located and analyzed.

●    Intelligent: according to system alarms, by docking ZTE’s VMAX system for big data analysis of network information and user behavior, alarms are automatically associated, root alarms are traced, and solution policies are matched to trigger related operations and reduce manual intervention, so as to handle alarms efficiently.

●    Closed loop: system resources are monitored, analyzed, policy-controlled, and re-monitored to keep the whole network stable and healthy.

Unified and Centralized O&M

After virtualization is introduced, software and hardware are layered by different vendors, and this increases O&M complexity. To solve this problem, ZTE introduced unified, centralized O&M to simplify virtualized network O&M. It achieves easy NFV O&M from deployment to optimization.

●    Centralized monitoring and layered management: system resources are centrally monitored. An orchestrator provides the global resource management view to monitor the configuration, alarms, and performance of physical and virtual resources. The OSS/BSS/EMS monitors resources at service layers of NEs. An independent maintenance interface is provided at each layer to meet operators’ layered management requirements.

●    Providing centralized maintenance tool and one-key management for network services: the deployment, capacity expansion, and service orchestration of virtualized network can be completed through one-key operation in the interface. It simplifies O&M in a virtualized system in the core network and shortens the system’s commissioning and capacity expansion period.

In October 2014, ZTE exclusively won the bid for China Mobile’s RCS convergence communications project, the largest platform and application project in the history of China Mobile. China Mobile’s active subscribers reached 16 million. This project shouldered the heavy responsibility of transforming China Mobile’s next-generation convergence communications strategy. It commercially deployed the virtualized IMS devices and RCS service platform for the first time, and it will become the world’s largest NFV commercial office and IMS-based RCS commercial network.