Building the AI era requires a new class of server infrastructure

Building the AI era requires a new class of server infrastructure

Artificial intelligence (AI) has leapt from the pages of science fiction to become a key component of organizations’ information management strategies.

According to IDC, worldwide revenue for AI software, hardware, and services will grow by 12.3 per cent this year to reach US$156.5 billion. That growth will accelerate as organizations recover from the COVID-19 crisis, with IDC’s Worldwide Semiannual Artificial Intelligence Tracker showing global revenue surpassing US$300 billion in 2024.

The phenomenal growth of AI has been enabled by massive increases in the volumes of data available for analysis, coupled with advances in processor speeds and improvements to AI algorithms.

But to lock these capabilities organizations need to invest in server infrastructure that delivers powerful general and AI computing power with intelligent operations and management capabilities. And this means looking beyond just the processing power of the machine and also considering the supporting systems such as AI accelerator cards, memory, storage, and power management to ensure they can meet enterprises’ requirements for diversified AI computing power.

When combined appropriately, these technologies enable the creation of AI systems that can derive insights quickly and enact decisions automatically. For example, the ability to rapidly identify the content of images and video streams has seen AI adopted for uses such as detecting security breaches in buildings, and helping doctors diagnose skin cancers. Being able to find patterns in huge volumes of data means AI is now commonly used to power automated financial trading, and to help retailers make better recommendations on goods and services for customers.

Essentially, any organization that wants to turn data into decisions in a rapid and automated way is building a capability in AI.

But unlocking this capability requires organizations to first invest in appropriate AI infrastructure, and that means properly understanding the speed and data requirements needed when selecting underlying infrastructure. Hence many organizations are revamping their data centers by implementing latest-generation servers with powerful general and AI computing power that can meet their requirements for processing the massive amounts of data required by AI algorithms. These servers utilize high performance hardware and specialized AI accelerator cards and support systems, including the ability to connect to high-speed and high-capacity memory and storage.

At the same time, these organizations also require intelligent operations and management (O&M) capabilities to appropriately manage their data center environments. Increasingly, that means implementing O&M that is itself powered by AI, to manage the increased complexity of these environments while allowing their operators to maintain or reduce operational expenditure.

All of these capabilities are included in the Huawei FusionServer Pro 2488H V6 intelligent server, which includes four physical Intel 3rd Generation Xeon processors, each of which contains 28 cores to make for a total of 112 cores. The FusionServer Pro 2488H V6 can incorporate Huawei-developed accelerator cards that boost the performance of storage, networks and AI computations, and supports 18TB of Intel Optane Persistent Memory, with flexible drive configurations supporting all solid-state drives.

This combination of high-speed processor, memory, AI accelerator card, intelligent NIC and storage ensures the FusionServer Pro 2488H V6 server has performance characteristics that are ideal for applications such as databases, cloud computing, virtualization, in-memory computing, and deep learning. In addition, Huawei servers incorporate five unique intelligent technologies covering maintenance, upgrading, energy saving, deployment, and discovery, which can help customers implement intelligent data center O&M and can reduce operating expenses by 15 percent.

Across the region organizations are working to upgrade their data centers with these intelligent new servers that not only meet the advanced requirements needed to power AI-driven systems, but which also lower power consumption and management overheads through intelligence operations and management.

These capabilities are proving essential for both meeting the demands of AI systems of today and for supporting future needs. As data volumes grow and the demands placed on AI increases, organizations will require the highest performance infrastructure if they are to realize the full benefits of AI-based insights and automation.

Deploying AI requires a platform that’s optimized for the workload. Intel’s 2nd generation Intel® Xeon® Scalable processors with Intel® Deep Learning Boost technology deliver scalable performance for end-to-end data science and a wide variety of AI applications, including machine and deep learning. Get started on the foundation for AI.