Chinese telecommunications giant Huawei is building out a total ecosystem and infrastructure play that brings AI-driven automation intelligence to every layer of the network fabric: from the core, to the communications pipe, to the cloud… and onward to every device’s chipset. Adrian Bridgwater was in London to hear its latest plans.
Arguably, Huawei (pronounced: wah-way) might find it tough to crack America in quite the same way that some other handset manufacturers have done, given the sensitivity of relations between China and the US, and the oft-cited security oversight of the company. (For example, the UK has a dedicated cyber security unit that monitors the company’s work as it expands its presence in the country – so far it has reported no national security concerns. – Ed.)
So Huawei probably has a much easier job on its hands when it comes to explaining how its technology roadmap and corporate development strategy is playing out, and why it is championing the use of automation and AI at every level of the network.
But above all, the company wants us to understand that it is not just a smartphone manufacturer.
Indeed, rotating Huawei CEO Guo Ping seemed to spend much of last year saying, “We’re so much more than a box seller.” (Huawei regularly changes its acting CEO to challenge the notion of a single person being in charge of a company.)
Whoever is in charge, the message has consistently been: Huawei is about cloud, networks, and services intelligence.
Significantly, the company has one of the largest R&D investment budgets on the planet. Huawei claims that it spends more than 10 per cent of revenues on R&D, and reports that 45 per cent of its employees work in researching and developing new products: that equates to 80,000 people worldwide.
Total ecosystem play
The ecosystem play for Huawei breaks down as a focus on: business partner alliances; its technology platform alliances (many of them open source); and the company’s infrastructure partnerships with various world governments.
But what Huawei wants to bring to the fore, in all its defined working zones, is a focus on automation efficiencies.
The this is not primarily automation in the sense of robots – although it is that too, in terms of home or industrial robot IoT applications; this is automation starting at the software layer. Huawei has been known to call it ‘digital operations’, but the technology application point is the same.
In other words, this is about automation brought about through applications being built more quickly, based on pre-defined reference architectures.
It is also about datasets that are more rapidly processed against policy-based rules.
And finally, it is about automation where data processing instructions have been replicated, based on templates drawn up from use cases within an industrial vertical or product group.
To put it simply, if one chunk of data management, analytics, processing, or even storage works well in one place, then it can be anonymised and used to build a similar structure elsewhere. This is software automation, and it has a massive application surface for the IoT.
Automated IoT use cases
Huawei’s Edward Fan, VP of carrier business group marketing, highlighted some real-world uses of this type of ‘automation blueprinting’ at a closed-door presentation in London this week. The meeting was designed as a preview of the company’s presence at the forthcoming Mobile World Congress in Barcelona in March.
Fan pointed to replicating IoT data use cases in specific smart-industrial and ‘smart life’ implementations. These include: smart street lighting; city sharing bicycles; livestock farming (to monitor milking and insemination); smart locks; and a range of other applications, including connected manhole covers, fire hydrants, intelligent post boxes, and aircon.
Many of these IoT products can now be activated by smart batteries to make make installation and configuration simpler.
Read more: Telia delivers smart mailboxes to Finnish postal service
Huawei will work with a particular focus on NB-IoT (NarrowBand IoT) deployments, and says that, globally, NB-IoT connections will reach 150 million this year. So far, the company has also undertaken a total of 1,000+ smart cities programmes worldwide, all of which employ some form of AI-driven automation.
Speaking at the same London event, Ryan Ding, Huawei executive director of the board, and president of the carrier business group, said that the carrier network itself needs to have AI controls and automation efficiencies engineered into its core – and into the services that run on it. This means that AI will exist at the cloud backend, as well as down the data pipe, and have a key role on the device itself.
Internet of business says
At the highest level, Huawei is simply capturing and defining best practices to compartmentalise service-based cloud automation opportunities: a key focus for much of the software industry. Yet in breakout Q&A press sessions at the event, more than one industry watcher tried to press Huawei on the technical mechanics of its approach to automation efficiency with little success. Does Huawei advocate empirical machine learning as the primary driver for AI on IoT machines, in post-production? Or does it believe that it is still too early to trust the machines – that we should really start from a policy-based control stance? Put another way, why is Huawei so tight lipped about what it thinks and believes? Perhaps the company is keeping more of its industry DNA to itself than it should. As ever with behemoths such as Huawei, its Western customers face a curious hybrid of free-market economics and Chinese organisational culture.