In this digital age, both connectivity and computing are the cornerstone for successful transformation. In a world where seven billion people and 100 billion of objects are intelligently connected, the diversity of devices is unprecedented. With more reliance of advanced technologies such as Artificial Intelligence, Internet of Things and so, a huge volume of data is being processed.
The nature of data thus evolves, and requires some unconventional ways to handle data. As such, further diversification of data has forced a paradigm shift in terms of computing power. Growth in data volume, meanwhile, has triggered a brute force to push from computing mechanism from simple scientific models to a predictive manner. A broader data distribution brings cloud to everywhere, where mobile data leads to more collaborations between all parties.
Evolving Data Structure and Complexity Requires Innovative Ways of Handling
Varying data structures along with its complexity, architectures. Diversified computing is replacing the single ones. Demand for computing power to speed up all workloads on processing also prompts the ever-growing need of diversified computing. And Huawei continues its innovations in both general and AI computing to respond to the challenges.
The tech giant has introduced the intelligent acceleration engine and the intelligent management engine for the FusionServer Pro servers . The series, ranging from rack servers, high-density servers, to two-socket servers and four-socket servers have been upgraded to fit the latest Intel processor to provide stronger computing power for customers, highlighted by the FusionServer Pro flagship models in 2288H V5 and 2488H V5.
These products are to address the challenges that many enterprises are now facing, as Forrester's Principal Analyst Danny Mu has discussed. Traditionally, data centres rely on CPUs to provide computing power. This has fallen out of fashion both in meeting the requirements of various services and delivering the desired economics. But heterogeneous chips can provide the best suited computing power for various workloads. Therefore, intelligent features become essential for offloading CPUs.
AI provides new approaches to improve the computing efficiency of data centres, via intelligent acceleration and management engines for instance. The intelligent acceleration engine improves server performance in terms of computing, storage, and network and offloads applications to hardware, thereby saving CPU resources and improving server performance from the system level. This is a giant step forward in transforming traditional data centres into intelligent ones.
The intelligent management engine, meanwhile, vastly boosts O&M efficiency for data centres. These two intelligent engines combine to shatter the boundaries of CPUs, servers, and L1/L2 of data centres. In addition to data centres, it is critical to provide high-performance, efficient, scenario-specific computing in various edge environments.
Diversified Computing Responds to the Challenges of New Data Landscape
This is right where diversified computing should get in. It refers to systems which use more than one kind of processor or cores to maximise performance or energy efficiency. The processors incorporate specialised processing capabilities to handle particular tasks. Diversified architectures break away from the traditional evolutionary processor design path, posing new challenges and opportunities in high-performance computing.
Our strategies include innovation, where Huawei invests in basic research and to improve computing power with innovative architecture in our chipsets. We also keep on our development in chipsets for both general and intelligent computing. Huawei is also providing our cloud services - instead of merely chipsets - for our clients and assisting them to further develop products which are best fit for them.
Ecosystem for Shared Success
We are indeed putting all these in action, particularly on the application level. For instance, the introduction of artificial intelligence, such as smart accelerators and intelligent management systems, helps improve the efficiency of computing in data centre. By carefully allocating tasks along different dimensions across the system, its computing power can be greatly boosted to shoot over the limits of the chips. AI can also help stabilising the system and smartly manage energy consumption.
Moreover, we bring a hyper-distribution model to work on a wide range of business scenarios. And this cannot be achieved merely by chips. Instead, we establish an infrastructure which can balance among computing, storage and connectivity by full utilisation of computing power, to break the performance bottleneck.
Such actions match well with the diversified computing strategy that Mu has described. And we share this common goal with our partners in the ecosystem. One successful case is GoVirtual, a high-performance computing solutions provider in Europe. Leveraging Huawei's solutions of diversified computing, the company develops autonomous driving system using virtual reality and noise, vibration and harshness simulations.
Computing power is one of the deciding factors of whether such complicated digital transformation is successful. We encourage enterprises to get fully equipped to go digital and get intelligent, where Huawei will leverage the 18-year experience plus unimaginable innovation to help them to further materialise their edges in this ever-changing intelligent world.