Now that a little time has passed since the Huawei Connect conference was held in Shanghai, it seemed like a good time to look back on a couple of the underlying trends from the event. The jumping off point for this is a round table Q&A with Guo Ping, one of the company’s rotating chairmen.
Three main themes emerged that are crucial to a number of wider issues related to what CIOs need to be thinking, particularly just how much the three main prongs of Huawei’s game plan they need to be considering. The chances are growing that they will soon need to address all three.
These are the need for a new computing platform, seen by the company as a prerequisite for addressing the other two, which are the coming of 5G communications and the combined issues and capabilities surrounding open source software. There is a fourth topic, of course, and that is on-going geo-political economic one. Until this year, it would be an obvious 'given’ factor that the USA would still be the dominant player in the direction and development of technology, both hardware and software. Others – China particularly – might lead in the building of systems, but the technology they used would be, predominantly, of US design.
Suddenly that is no longer the case, and 5G – plus the devices that will exploit it – is the key driver. As Guo Ping pointed out, the number of countries developing local applications for 5G is growing, and countries like Japan, South Korea, France, Russia and of course China itself not only provide a substantial and already energetic 5G market, but that market is happening in both consumer and business/industrial sectors. Recent evidence now suggests that the UK will also be amongst those players.
The key thing here, of course, is that 5G is about much more than even smarter smart phones for trendy millennials and GenX-ers. Indeed, that market is likely to be pretty secondary. It is the many different business sectors that are going to drive the growth and development, and this is where CIOs need to be paying particular attention. With 5G communications providing the possibility of flexible, movable, and above all very fat pipes – working with data where it is most needed becomes not just possible but essential to business.
Data centers will dissolve and become virtualized across corporate networks, including public cloud operations, and the volume and quality of data will be such that existing systems and processor architectures are likely to be stretched to – and possibly beyond - breaking point. This is particularly the case when the rapid growth in new AI applications is thrown into the pot. Apart from anything else these are best run on processors with a fundamentally different architecture from the venerable – and rightly venerated – general purpose x86 processor family. That, as Guo Ping made clear, is one of the key targets that Huawei has in its sights.
With a global marketplace he reckons as worth $2 trillion, he is understandably keen that as many customers as possible can get to build their computing capabilities on top of that primary connectivity business. To help that happen, it is therefore being quite specific and controlled in choosing which elements it needs to provide and which to leave to the cjustomeers. For example, it intends to focus on the processor cards and associated components based on its Kunpeng and Ascend processors and help system providers develop services and solutions for its customers.
Guo Ping was particularly keen to point out that Huawei now sees an opportunity for these processors as alternatives to the established device families, not least because of the perceived weakness of of the latter at running AI applications, particularly as they become more complex:
We aspire to provide a processor cards computing platform as an alternative for the world. I think this is an important solution that the businesses from the UK and other countries can look to as they seek business continuity and a plan B for heterogeneous computing. And before we officially put the computing platform on the market, we had already been using it on more than 100,000 of our servers within our company. So, it's already a mature technology.
One of the interesting undercurrents here is the company’s commitment to open source software, both as a user and a contributor. Even when it does set out to develop significantly new lines of software, as it did earlier this year with the introduction of HamornyOS, open source contributions from many others play an important part.
For example, though Hauwei’s primry contribution to 5G development is in the communications infrastructure, it is well aware that much of the reason for anyone using it will come from the applications that are developed as a consequence of its existence. Those applications will come from around the world, and some will certainly have global impact over time. Being as open as possible with the software infrastructure, therefore, gives those developers the best possible chance to flourish.
Guo turned to software history to make this point for him:
Essentially, open source is just one of the business models that leads to business success. A case in point is the competition between Apple II and IBM's PC. IBM chose the open-source model, which resulted in the wide-spread success of today's PCs. In its competition with iOS, Android also opted for an open-source model, which has allowed more vendors to come on-board. As a result, Android has captured nearly 85% of the market share.
The reason why we adopted an open-source approach for our computing platform is that we hope we can attract more vendors and users, and this way, they can benefit more and achieve business success as they deploy AI and computing platforms. Huawei is willing to make more contributions for this reason.
And of course the goal here is not just to create more opportunities to sell 5G switches, antenna and other infrastructure components. With its new ARM-based processor families and server systems that run on them, the company is now moving up the 5G applications food chain, particularly where Artificial Intelligence and Machine Learning applications are concerned. These are most certainly going to be based on open source contributions from around the world, and it is reasonable to guess that many of the developments and breakthroughs in these fields will come from non-western sources, including nations where China in general and Huawei in particular already have a strong foothold:
We have made our systems open source, because we believe open-source systems are the most competitive. I'm very pleased to see that many open-source organisations have moved their headquarters to permanently neutral states like Switzerland. I believe this is a trend that will spread to more open-source organisations, as they want their systems to be widely used by the world's seven billion people.
One of the inevitabilities with 'the natural order of things’ is that the order always ends up getting disrupted. One of the key natural orders in the computer industry and wider IT domains has been that the key developments come out of Europe and the USA (with the latter being largely European by proxy). But over the last 10 years or so (much longer if one wants to include the development of semiconductor technologies used to make the essential computer chips) the lead has changed hands. Many of the leading companies, though still based in the dominant IT marketplace of the USA, are the product of Indian, Korean and Chinese minds.
With Huawei being one of the dominant developers and providers of the infrastructure underpinning the next developments – 5G, AI, ML and computing out to the edge – and Korean SK Telecom becoming a leader in 5G systems implementations there are signs that the centre of gravity for computing and IT is moving eastwards.