Most consulting organizations have predicted that NFV will achieve large-scale commercialization in 2020, such as AT & T hopes to achieve 75% of network virtualization in 2020. 2017 will be a critical year accelerating deployment of a large number of commercial cases. Leading operators such as VDF, speed up SDN / NFV pace, and will shorten the transition strategy from 5 years to 2 years, by building a unified cloud platform in advance for future carrying diversified business covering existing and new innovations.

"The NFV implemented the standardization and virtualization of ICT network equipments, but in software architecture and operating model, it still followed the traditional way," said Huawei CEO Xu Zhijun, who first raised the concept of NFC at the analyst conference in 2016. " We should step forward with the concept of cloud, moving the network software into a ‘fully distributed, fully automated’ environment. This is the real meaning of the equipment cloudification (NFC). "Xu Zhijun also stressed that Huawei is committed to the introduction of cloud technology and extension to the basic network of various fields.

The rooted reason of operators practicing NFV, used to be cost control due to vendors owned the proprietary of dedicated and non-standardized hardware. Now the business has become a new driver, especially in the coming 5G era, VR, AR, AI and IoT will significantly impact business forms as well as user behavior, which is bringing new harsh requirements to the traditional telecom network. Users are requesting extreme experience with lower and lower tolerance to the latency. Business like IOV (Internet of Vehicle), real-time data transmission in car crash tests, would request the network latency down from milliseconds to microseconds or even nanosecond level.

With the continuous expansion of such business, telecom operators are bound to reshape the network architecture, and build a distributed, multi-layer DC architecture with different business deployment principles according to various latency and reliability needs. For example, the supporting systems like BSS, OSS, etc are not latency-sensitive, and can be centralized in core DC with DataCenter disaster recovery design. CDN, video or IoT business should be deployed in access DCs to be closer to the users and hence ensure customer experience.

To build a telecom cloud architecture, there are many real challenges need to be considered and tackled,

First, how to ensure the Cloud platform has no interoperability issue, since current telecom network are composed of multiple network functions from a dozens of different manufacturers. The deployment project numbers and interoperability testing proofs are important aspects to verify if a cloud platform is capable of. The traditional CT vendors like Huawei, till the first quarter of 2017, have implemented more than 260 carrier NFV projects in the globe, and its fusionsphere cloud platform has gained testing proofs with over 20 VNF / SDN vendors, which can be strong proofs to let carriers assured of the platform capabilities. In the openness of the platform, although OpenStack has established a standard platform, the commercial products by multi-vendors based on open-source codes may not be able to interoperate with each other. As of fourth quarter of 2016, OpenStack community harvested a big tent project "TriCircle", which is contributed and dominated by Huawei, aiming to drive the openness among OpenStack players.

Second, cloud must adapt to diversified workloads especially in the field of telecommunications. High forwarding, high throughput, low latency and five-9 or six-9 high reliability, have become the tags of special requirements to telecom networks and telecom IT. In terms of performance, virtualization is bound to consume more computing resources and lower overall performance. How to keep the high performance is a critical subject. The cloud should have to be able to provide different kinds of resources including physical resources as well as virtualized resources, to satisfy different business needs. At present, Huawei has been able to provide different IT products and solutions to accelerate the performance in aspects of computing, network and virtualization, such as providing FPGA plug-in, to offload the computing workload of DPI, 4G paging and SIP protocol Resolution, which can be improved by 40% -70%; in virtualization layer , high-precision VMs and DPDK-accelerated OVS, will significantly speed up the high forwarding by 50%; in network layer, Huawei can provide 100GE NIC, iNIC and 10Gb VM to greatly meet the high demand for network forwarding like vEPC.

Since the IT system are decoupled comparing to traditional closed and dedicated systems, it brings new reliability issue of how to ensure the high availability of each laye and still ensure five 9 or six 9 carrier-class system reliability. Huawei has invested a lot on DC level and system level reliability design, and provide active-standby and active-active disaster recovery solutions, especially based on OpenStack. In 2017 there will be delivered in several NFVI projects. In the meantime, a hierarchical solution can also be in place to ensure every layer has its availability design, for example, in VM layer, OpenStack management control nodes HA without a single point of failure, VM live migration, VM one-second fault detection.

Third, the Cloud should not only meet the basic requirements of the business, but also unleash its basic “As-a-service and automated”functions to create business value, like bring efficiency to operations, and bring agility to business evolving. For example, it should continue introducing and adopt latest technologies such as container, to enable deployment in seconds, and enable faster TTM of digital business. As to development environment, digital business like IoT demands a release refreshment every 2-3 days per application. Deploy a streamlined devops based environment is very necessary. Manual operations and repeatable configurations in every step, won’t work in this condition. Huawei has already adopted container in R&D environment with million-level container management capability, and deployed them across DCs to support a automated process.

We are already in Cloud era. NFC will replace NFV as a new term for network transformation. Plan ahead, to be future ready.


Knowledge Network Articles enable our partners to create high value content to express a particular view or highlight an opinion in an editorial style environment on Total Telecom, they are not produced by our editors and do not reflect the views of Total Telecom. To find out about content creation opportunities for your organisation visit the information page or email