The world is witnessing massive growth in data centre computing as organisations build the capacity needed to store and process huge volumes of data, with Gartner predicting growth in the large data centre category of 3.6% CAGR.
While the public cloud is proving popular, many organisations are seeking to retain control of their data and computing environments by hosting in shared private facilities or creating their own private data centres.
But the need to meet highest standards of performance and security means the technology chosen to operate these new centres is critical, especially the switching infrastructure that forms its communications backbone.
The need for AI
Digitalisation is a key factor behind data centre growth, and many are being built to meet the specific requirements of modern systems, including those that incorporate artificial intelligence (AI). These AI-based systems demand incredibly high data throughput and have great sensitivity to packet loss; studies show that the loss of even 0.1% of packets can halve AI computing power.
The sophistication of these modern environments also requires a high level of specificity in terms of data they generate regarding their own performance, to assist in the rapid identification and rectification of problems. And of course, considerations such as the cost of power and cooling remain an ongoing concern.
All these considerations generate complexity well beyond those traditionally encountered within data centres – and demands a new approach to their management and maintenance.
The answer lies in utilising AI itself to play a key role in data centre operation and maintenance. When embedded in the operating environment of a data centre, AI can take on a large component of monitoring, management, and optimisation duties, including the automation of low-level administration tasks and fault resolution, while providing decision support to human operators when required. This leads to more reliable and more efficient performance and faster fault resolution.
The CloudEngine 16800 advantage
Huawei is a long-time leader in the data centre switch market, having sold more than 32,000 core switches to clients including the Max Planck Institute in Germany, China UnionPay, and ICBC Asia. And its latest flagship CloudEngine 16800 series data centre switch is an AI-driven switch the meets the needs of the AI-driven era.
The CloudEngine 16800 features real-time learning and training capabilities in addition to autosensing and auto-optimisation of traffic to enable exceptionally low latency. Millisecond-speed data collection across eight dimensions provides complete network status data aggregation for efficient operations and maintenance, while intelligent analysis for self-optimisation identifies faults in seconds. It also uses an industry-unique AI algorithm, iLossless, to achieve zero packet loss on the entire network, to ensure that AI-powered workloads can perform at peak efficiency.
The CloudEngine 16800 is also the industry’s highest density 400GE switch, with a switching capacity up to five times the industry average. This future-proofs data centre investments and satisfies foreseeable requirements for smooth network architecture for the next 5-10 years.
Additional capabilities include ultra-high-speed signal transmission, super heat dissipation, and an efficient power supply, which together reduce per-bit power consumption by 26%.
It was this combination of features that stood out to judges at Japan’s largest and most influential ICT exhibition, Interop Tokyo 2020, where the CloudEngine won the Best of Show Award Grand Prize.
With data centre performance becoming more and more critical to business performance, organisations need to make the right investments.
By incorporating AI technology into the industry’s highest-density switch, the CloudEngine 16800 ensures organisations can make the right investment both for today, and well into the future.
To find out more about CloudEngine and other Huawei’s offerings, visit here.