Times seem to be rolling on at blazing-fast speeds after we enter the Internet era. Looking back at the technology roadmap of the past decade, one would not be surprised to find that it is marked by a long string of fast-changing technologies, among them the cloud technology, big data, artificial intelligence (AI), and blockchain, just to name a few. The rapidly emerging technologies have been constantly reshaping our realms of knowledge, and have brought unprecedented challenges as well.

The past decade of the Internet has been strewn with countless new technologies, where novice business models were brought to the consumer market one after another. Looking ahead, the next decade will be one of convergence bolstered by technologies. The enterprise market will witness more rapid changes and finally embrace digital transformation.

According to the Global Industry Vision (GIV) 2025 report released by Huawei, the total number of global connections will reach 100 billion by 2025, and 86% of enterprises will use AI for their business, creating a digital economy valued at USD23 trillion. This will be accompanied by explosive volumes of demands for intelligent applications by the whole society. In this context, intelligent computing is poised to become a new infrastructure.

Qiu Long, President of Huawei Intelligent Computing Product Line, said, “Enterprise digital transformation is moving towards the intelligent direction. The following three characteristics have been demonstrated: First, traditional enterprises are shifting from general-purpose servers to intelligent servers. Second, digital transformation is shifting from closed to open platforms. Third, the underlying hardware enabling enterprise intelligence is shifting from monolithic to heterogeneous and then to the edge.” Therefore, Huawei announced its intelligent computing business initiative based on the Boundless Computing strategy on May 31, and launched the KunLun V5 server and ES3000 V5 series.

Huawei Has Been Laying the Groundwork for Intelligent Computing

On the one hand, AI has been the keyword resounding throughout the major meetings of technology giants in 2017. For example, Alibaba touted the ET Brain at its Cloud Computing Conference. At the Baidu AI, Big Data, Cloud Summit (ABC Summit), Baidu debuted autonomous driving technology that is ready to hit the road in Beijing. Another company of the China Internet Big Three, Tencent, announced that it will open AI technologies in computer vision, intelligent speech recognition, natural language processing, and other fields. On the other hand, from the perspective of intelligent computing, Inspur and Lenovo have launched their intelligent computing proposition. So, is Huawei late for the intelligent computing party?

Of course not. Firstly, in fact, as Huawei has made it this far, one rule of thumb has been Huawei’s mainstay: never lose sight of customers’ requirements. Huawei has constantly stayed true to this motto.

The first wave of AI was unleashed after the AlphaGo beat the world champion for chess in 2016. Therefore, 2017 was virtually the first year for AI to gain public attention on a global scale. The same year, Huawei first put forward the vision to build a fully connected, intelligent world, which signals its strategic shift to intelligence on a company scale.

Secondly, Huawei’s IT Product Line has long ago proposed the idea of heralding in an intelligent future with smart IT. This concept is essentially to leverage intelligent technologies to fully explore the potential of infrastructure. For example, to bake intelligent capabilities into hardware resource pooling, infrastructure cloudification, and data. Doing so is to implement smart IT and drive enterprise business intelligence, enabling smart IT to become the driving force for business value creation.

Thirdly, all of Huawei’s intelligent innovations are centered around customer requirements. Therefore, Huawei’s interpretation of intelligence is not solely a proposition on its own. Smart IT is a means rather than an end. The goals are to enable a smart future and smart industries. Specifically, to use smart IT to promote smart government, smart communications, smart transportation, smart finance, and smart retailing. This is the future that Huawei looks to realize through intelligent computing.

In particular, in terms of product transformation, Qiu Long said that Huawei provides intelligent engines that enable servers to think, all the way from intelligent management, diagnostics, deployment, and mobile management, to asset management. The intelligent engines take servers to a whole new level of intelligence.

Therefore, Huawei is by no means a late comer to the intelligent computing party. Huawei is not a me-too vendor of intelligent server products and solutions, either. Instead, Huawei keeps customer and industry requirements in mind, drills down deeper into the common demands, and strikes a well calculated move based on deliberate decision. Every punch is dealt precisely.

Why Has Huawei Upped the Intelligent Computing Game?

In my recent talks with some enterprise CIOs, I’ve heard more than one time that traditional enterprises’ requirements for AI computing are exploding. In the past, Internet companies are the major group investing in and using AI applications. Nowadays, as the digital world is transforming into an intelligent world and industries are pushing deeper into +Intelligent, AI computing capabilities are becoming the key differentiator for enterprises in the future.

At the same time, AI is hatching new business opportunities in many new industries, such as autonomous driving and smart home. Once these industries mature, new requirements will rise for intelligent IT infrastructure. This is the core reason why Huawei establishes the intelligent computing business initiative.

In addition, there are two underlying reasons from the technical perspective: Global data center construction is increasing both in speeds and scale. With the blazing-fast development of services, IT infrastructure needs to be quickly deployed, brought online, and conveniently managed. The management scenarios of massive servers will become more and more complex, and the traditional O&M faces many new challenges. Intelligent energy consumption and fault management capabilities can better guarantee the Service Level Agreement (SLA) and continuously save IT infrastructure OPEX by as high as 20%. Another reason is that the traditional x86 architecture based on CPUs is no longer the optimal choice for handling massive computing requirements of applications such as AI. Instead, heterogeneous computing based on various computing units such as GPU, FPGA, and ASIC is emerging as a potent computing force to deliver rapid intelligent development.

In my opinion, Huawei has been keen in discovering the change of requirements. Therefore, Huawei has built an all-stack AI intelligent cloud hardware platform to make the data center more intelligent. This is the Atlas intelligent cloud hardware platform launched last September.

Epitomizing the very essence of Huawei’s high-performance heterogeneous computing servers, Atlas uses heterogeneous computing resources such as GPUs and FPGAs, improving AI performance by 10x over traditional x86 servers. In addition, the intelligent scheduling platform provides unified orchestration and scheduling, decouples GPU, HDD, and SSD resources, provides hardware resources on demand, and implements deployment and logical server provisioning in seconds, greatly reducing the service deployment cycle.

Thanks to Atlas, Huawei has been able to implement intelligent computing in many specific application scenarios. In March 2018, Huawei and YITU jointly released a panoramic, ultra-precise facial recognition solution engineered for safe city and smart campus. The solution implements over 99% identification accuracy and provides up to 256 channels of real-time analysis.

The Shenzhen Traffic Police Corps has built an “urban traffic brain” based on Huawei’s G series heterogeneous servers. The solution has improved the road capacity of Shenzhen by about 8% and delivered a 10x efficiency in traffic violation image recognition to enable more efficient, intelligent law enforcement. The project won the Safe City Award at Smart City Expo World Congress 2017 for its innovative practical and social significance.

When Talking About the Intelligent Computing Transformation, Huawei Means It

Objectively speaking, the announcement of Huawei’s intelligent computing business initiative is made just at a perfect timing.

(Translated from the Article by the KOL Zheng Kai)

The post Huawei Ups the Intelligent Computing Game appeared first on Huawei Enterprise Blog.

Source: Huawei Enterprise Blog
—