Written by Ken Wieland, contributing editor at Light reading
The telecoms industry faces a huge structural challenge. Opex, according to various market estimates, has been increasing faster than revenue for over a decade. Spiralling O&M costs is a big part of the problem. There is the challenge of meeting increasing demand for services among individuals and enterprises, as more radio access technologies (RATs), frequency bands, and sites are being deployed with differential capabilities. Moreover, how do you make network energy consumption stable while traffic may boom 100 times in the following 10 years? It’s not all doom and gloom, however. Increased network automation can ease the financial strain.
As part of its ongoing efforts to help industry achieve the goal of autonomous networks, Huawei, in the run-up to this year’s Mobile World Congress (MWC), officially launched its innovative IntelligentRAN architecture. It is designed to reduced O&M costs of wireless networks through greater automation, as well as optimise user experiences and meet SLA requirements. Increased energy efficiency is another goal.
“IntelligentRAN is a major breakthrough,” asserted Calvin Zhao, vice president of Huawei’s Wireless SingleOSS product line.
Speaking to Light Reading on Huawei’s stand at MWC, Zhao explained that IntelligentRAN builds on the launch of Huawei’s Autonomous Driving Network concept in 2018.
“With our efforts and practices over the last few years, we have always been looking at use cases that bring value to our customers,” added Zhao. “These efforts include our proposed IntelligentRAN concept, which is an implementation of the ADN framework.”
ADN details various ‘Levels’ of automation from L1 to L5 (a fully autonomous network). L1 to L4 describe an ever-increasing reliance on automation (and less reliance on the need for highly skilled personnel) to monitor and intervene. IntelligentRAN, as far as Zhao is concerned, is an ADN booster and a way for carriers to advance up the Level hierarchy.
Zhao explained that IntelligentRAN was developed to tackle three areas of vital concern to mobile network operators: agile service provisioning, network optimization, and O&M cost reduction. These three challenges have become all the more pressing with the arrival of 5G.
On agile service provisioning, observed Zhao, 5G poses extra difficulties in that there are potentially many more services with the next-gen tech than 3G and 4G. Moreover, these services will have different network requirements, which makes service-provisioning complicated. “That’s why, in the 5G era, we’re thinking about flexibly combining different services in different frequency bands for optimal network performance potential,” he said. “This we believe brings value to our customers.”
Huawei is also looking to leverage IntelligentRAN for multi-dimension coordination for maximizing network performance potential and bringing optimal user experience in the context of multi band and multi site heterogeneous networks. “Services are moving between cells, so there has to be a capability to forecast that and to ensure handover and improve service consistency,” asserted Zhao. “This will help improve user experience.”
Moreover, in the 5G era, there is a danger that opex costs spiral out of control. Huawei notes that 5G networks might generate as much one million alarms per day. “We hope to have the capability to filter out the root cause efficiently and ensure both accurate and effective work order,” said Zhao. “This will reduce cost in the assignment of work orders and simplify the whole process. In terms of O&M we hope to ensure a reliable, always-on wireless network with zero interruptions.”
The IntelligentRAN architecture, like the ADN concept, is hierarchical with domain-based autonomy and allows vertical cross-domain collaboration. Base station sites sit at the bottom layer with the network layer directly above. On top of the network layer are platforms and applications.
The ‘brain’ of IntelligentRAN is the Mobile Internet Engine (MIE). It’s designed to better coordinate data, models, and decisions between all layers. In this way, says Huawei, a path towards wireless intelligence can be achieved.
“The most difficult layer [to bring intelligence] is the base station sites, but we will have MIE-RT [real-time],” said Zhao. He lauded MIE-RT as a breakthrough since sites could now “train” themselves, which is important in lowering power consumption and optimizing the user experience.
“Another major MIE breakthrough is prevention and forecast capabilities for network faults, which is a key L4 feature,” added Zhao. He explained that Huawei’s past approach to alarms was to classify them and apply algorithms. By combining the alarms together, as part of the diagnosis process, Huawei could then provide the root cause of the problem.
“This was the past approach,” emphasized Zhao. “Based on historical data we can now predict the potential issues that may happen on components and network performance. In this way we can be more proactive rather than being passive.”
Zhao flagged too that IntelligentRAN can be seen as an upgrade to Huawei’s SingleRAN solution, which he claimed had already proven hugely successful in reducing deployment costs through a single module deployment approach that can support different ‘Gs’.
“We’ve now introduced intelligence into this fundamentally simple SingleRAN architecture we’ve already built,” said Zhao. “The whole idea is to introduce intelligence capabilities – including intelligent algorithms, big data capabilities – to make the network more intelligent to support such things as as-a-service provisioning, zero network faults and optimized experience.”
Moving forward, but challenges ahead
Since announcement of ADN in 2018, Zhao said Huawei has been working with carrier partners to put the concept into practice by exploring use-case examples to introduce intelligent algorithms into different network domains, such as wireless, optical, fixed and cloud.
“We’ve also defined intelligence for these different areas. but the focus so far has been mainly on O&M and opex reduction,” admitted Zhao. “However, we have enhanced our cooperation with top carriers – including China Mobile, Vodafone, Deutsche Telekom and LG U+ – to identify more use cases.”
One of those use cases, he said, is reducing energy consumption in the shape of Huawei’s PowerStar solution. “In China this product has helped customer save 10% of energy.”
Yet autonomous networks, pointedly noted Zhao, are “still a direction without a complete standard.” Because of the huge amounts of data involved, he called on industry to work together on “defining data standards and formats so we can achieve coordination between upper and site layers.”