Huawei recently released Atlas, a new-generation, intelligent cloud hardware platform designed for the public cloud, AI, and HPC scenarios.
Positioned for scenarios like public cloud, artificial intelligence (AI), and high-performance computing (HPC), Atlas rides on key technologies like heterogeneous resource pooling and intelligent orchestration to deliver new levels of hardware resource utilisation and adapt more flexibly to service demands. As a result of Huawei’s Boundless Computing strategy, the Atlas platform is intended to position Huawei to steer the infrastructure resource pool revolution.
With the rise of AI research and application, AI and cloud computing technologies will be integral to people’s daily lives. From applications like smart manufacturing to autonomous driving, AI and cloud technologies are spreading. Meanwhile, pivotal to intelligence is having capabilities that allow numerous systems to process in real time the massive volumes of data generated from a world of full connectivity.
For example, a city with 100 000 cameras installed can produce 100 billion car plate records and 1 trillion facial recognition records a year. It has become a tremendous challenge for traditional computing systems to deliver the required capabilities in order to process, query, and analyze data of such a high order of magnitude.
Built for the public cloud, AI, and HPC scenarios, the Atlas platform addresses the data processing challenges in the AI era. The Atlas platform is based on Huawei’s FusionServer G series heterogeneous servers. The FusionServer G series leverages technologies such as heterogeneous resource pooling and intelligent orchestration to pool resources like GPUs, HDDs, and SSDs, and provisions hardware resources on demand to suit the needs of specific service models.
Huawei provided the following information:
Atlas drives an over 50% higher resource utilization efficiency, and boosts performance by over 10x compared with the traditional x86 architecture, meanwhile slashing the hardware device types for customers. Additionally, Atlas can provision logical servers in different resource configuration ratios, and deploy the servers within seconds, significantly shortening the service rollout cycle while empowering customers with the flexibility to adapt to service changes.
“As the AI era is approaching, traditional hardware cannot meet the requirements on development of AI and cloud technologies.” said Qiu Long, President, IT Server Product Line, Huawei. “Positioned as a new-generation intelligent cloud hardware platform, Huawei’s Atlas pioneers heterogeneous resource pooling and intelligent orchestration technologies to bring resource utilization and performance to new high levels. Atlas is also the fruit of Huawei’s Boundless Computing strategy. Huawei has been inspired by the idea of making computing simple through continuous innovation, and keeps customers’ requirements in mind to build servers that are stable, reliable, and high-performing, to help customers better cope with the challenges for successful transformation in the AI and cloud era.”
Huawei has strategized Boundless Computing as a major step to building a better connected world. The Boundless Computing strategy advocates innovations around computing and leverages innovative chips, heterogeneous computing, and edge computing in order to push the traditional computing boundaries that used to be confined to CPUs, servers, and data centers. At HUAWEI CONNECT 2017, Huawei has released multiple brand new public cloud services based on the Atlas hardware platform, including GPU and FPGA cloud services, empowering AI applications with cloud capabilities and enabling customers to access heterogeneous computing services with more ease and speed.