PARTNER FEATURE: Automation is 5G’s fourth fundamental dimension, and will be a key driver of 5G business success.
Operators worldwide started deploying 5G networks last year, and 5G’s development is now in full swing. It took 10 years before there were 500 million 3G subscribers and 5 years for 4G subscribers to reach that number. It is estimated that it will only take three years for 5G to attract 500 million subscribers owing to the unprecedentedly simplified mobile networks.

5G has been built on three fundamental dimension, including enhanced mobile broadband (eMBB), ultra-reliable and low-latency communications (URLLC), and massive machine-type communications (mMTC). The mobile industry’s technology, application, and business innovations over the next decade will center on these three key application scenarios, and will contribute to a prosperous 5G era.

However, the introduction of 5G technologies will significantly increase network complexity, resulting in rising OPEX for network operators. Hence, developing sustainable OPEX models is now becoming urgent. Operators have to find an innovative way to stop the OPEX mounting while networks are becoming more and more complex. The key solution is to make the network automated, which we consider as the fourth dimension of 5G in addition to the other 3 fundamental dimensions.

Network automation could not only help operators keep their OPEX in check as the network complexity increases, but also encourage further digital service innovations. To embrace the digital market, operators need to be as agile as over-the-top (OTT) service providers in capturing business opportunities. When the networks are automated, service time-to-market (TTM) will be accelerated and more business opportunities will be opened up.

Diagram of 5G’s capabilities: triangle to diamond

Automating mobile network in the 5G era has been becoming a consensus in the industry. And the industry is now still at the early stage of network automation. How the industry can keep pushing the limits of network automation is still an important question. A fully automated network cannot be achieved through a single product’s innovation, but will only come about through system architectures innovation and use case innovation based on workflow scenarios.

Breakthrough in automation capability by architectural innovation

For the system architecture innovation, we take the design of autonomous driving cars as an example, of which we think the conception is basically the same as that of automated networks. The existing cars, originally designed for human operations, have complicated systems requiring frequent human interaction. When the vehicles are gradually automated, and eventually rely on themselves, the system architecture of vehicle should be redefined, and each vehicle component should be redesigned as well with automatic driving capability.

Automation of telecommunications networks is also challenging, similar to vehicle automation, but network automation is even more difficult because of the complexity of service management. One reason contributing to this complexity is that mobile networks are dynamic, as traffic models and wireless channel environments change constantly. In addition, wireless networks involve thousands of parameters, and the interdependencies between these parameters challenge the most experienced industry experts.

Essentially, network automation relies on an automatic end-to-end (E2E) closed-loop intent or policy. This requires a shift in system interaction between network layer and upper layer (unbrella system) from a simple exchange of data/parameters to an exchange of policies/intentions. This shift, in turn, requires mobile networks to have intra-domain autonomy, that is, autonomy by layer and coordination between layers with openness.

Autonomy by layer and coordination between layers with openness

Service-level automation through autonomy by layer and coordination between layers with openness can be achieved by leveraging unique advantages of operators and vendors. Operators define and orchestrate automation workflows. From a network perspective, vendors is able to efficiently find solutions to address network complexity and create new level of simplicity, and by doing this to unleash network potential. Vendors can also help to implement policy-based and/or intent-driven network autonomy by providing standard open scenario-based interfaces.

Huawei launched the “Cloud AI – Network AI – Site AI” three-layer architecture in 2018. This architecture is aimed at creating “autonomous driving” mobile networks, and contributes to the realization of autonomy by layer and coordination between layers with openness. The three-layer architecture deeply integrates artificial intelligence (AI) technologies, and this will build automation and AI capabilities in all layers including cloud, network, and site.

The hierarchical autonomy of Huawei’s “Cloud AI – Network AI – Site AI” network architecture allows each network layer to fully unleash its automation potential. The architecture also creates efficient closed-loops between network AI and cloud AI, as well as between site AI and network AI. As such, Huawei’s three-layer architecture lays the foundation for creating automated networks in all scenarios.

Automated use case innovation based on workflow scenario

Network architecture is the foundation of network automation, but whether or not networks are successfully automated depends on finding scenario-oriented use cases for automated networks and innovating based on these use cases.

The mobile network industry started exploring automation at a very early stage of the mobile broadband (MBB) era. First focus on tool-assisted automation and then function-based automation. We believe that network automation should be scenario-oriented, and create value for the operators through closed-loop automation systems.

Huawei unveiled its Mobile Broadband Automation Engine (MAE) at the pre-briefing meeting for Mobile World Congress 2019 in London. MAE is a wireless network automation solution providing scenario-based network automation in five areas—network planning, deployment, maintenance, optimization, and service provisioning. Besides, to meet the requirement of E2E service automation, MAE will provide scenario-based APIs as an open platform to enable intent driven network autonomy.

Take site deployment as an example, MAE provides a zero-touch site deployment solution, with automated site design, on-site deployment, and site verification. This solution will effectively shorten the time needed for deploying 5G sites.

Huawei’s MAE has been tested across multiple pilot projects globally in North America, Japan, Korea, and other regions. An operator in North America has used MAE to reduce the manual labor for site deployment. The result is very encouraging. Site deployment duration has been reduced from four hours to half an hour per site.

WTTx Suite, another Huawei solution, realizes the automated network service provisioning. It offers agile and reliable fixed wireless access (FWA) adapting to operators’ business and operation support system (BOSS), network planning, network optimization, and customer service systems. More than 50 operators currently have deployed WTTx Suite for their 4G networks all over the world. This has laid a solid foundation for 5G development and will speed up the service provisioning of FWA as the first commercial 5G business.

It will take some time to achieve ultimate mobile networks automation. To realize network automation, the industry needs to jointly define clear standards, encourage technological innovation, and steer implementation.

Huawei is committed to working with the industry partners to build automation capability in wireless network, and looks forward to creating an intelligent and fully connected world with all industry players.