Huawei HDC Cloud 2021: Details of the cloud-native and AI technologies announced at the event

On April 25, 2021, Huawei began the HDC Cloud 2021 conference and announces the 6 revolutionary products for the cloud-native and AI technologies at the Huawei Developer Cloud Conference 2021 in Shenzhen. At the same time, the company also announces a US$220 million investment into its Huawei Developer Program 2.0 this year.

These technologies will deliver refresh algorithms and quality, which helps them to stretch their limits and produced more quality products. This new technology lineup includes-

1. Huawei Cloud CCE Turbo cloud container cluster
2. CloudIDE intelligent programming assistant
3. GaussDB(for openGauss) database
4. Trusted Intelligent Computing Service (TICS)
5. Pangu Model (including the world’s largest Chinese NLP model and CV model)
6.  infrastructure software for diversified computing

This will jointly bring us towards the more enhanced cloud and AI industry. Let me take you to explore all these techniques in detail-

Huawei Cloud CCE Turbo cloud container cluster:

This new cloud CCE Turbo focuses on high-stream computing, server networks, and scheduling because it revolves around the enterprise core systems, innovative services, container performance, elasticity, and scheduling capabilities. The three aspects of the CCE Turbo are as follows-

1. Computing acceleration: It’s the only container cluster in the industry that supports full container component offload with no loss of server resources or performance. So it’s good enough to help VIPKid, which is a leading education company tends to improve 40% audio and video service performance.
2. Network acceleration: This cluster cross the network transforms a two-layer network into a single-layer network. It helps Sina Corp, which is a leading social media company in China suffering from traffic surges.
3. Scheduling acceleration: It senses different application models and network topologies, automatically optimizes task scheduling policies. Simenntenoiusly implements large-scale hybrid deployment and intelligent scheduling of offline and online services.

You can see 40% more results in all three above-mentioned characteristics.

Huawei CloudIDE:

This is a smart programming assistant for developers and delivers a lightweight and fast programming experience. It also provides the developer the intelligent online programming performance anytime, anywhere, and on any terminal. It covers the following aspects:

1. It can find all completion results and determine and sort possible results, improving the accuracy of multi-symbol and long results.
2. It supports completion for complex contexts and even obscure third-party APIs.
3. By automatically avoiding recommending outdated, obsolete, and dangerous APIs makes your code more robust and secure.

GaussDB(for openGauss) database:

This database will be proven to be a trusted choice for migrating enterprise core data to the cloud. Now, an enterprise can freely switch to the cloud and become data-driven. We can define this GaussDB in the following aspects-

1. High performance: It works with 54% higher transaction capacity in comparing to the existing products in the industry, and its complex query latency is 82% lower than that of similar products.
2. High elasticity: In aspects of the traffic peaks, the computing nodes can be expanded within minutes whereas the torage capacity in seconds. Furthermore, the services are not affected by this entire equation.
3. High availability: When a single point of failure occurs, the unique Switch Turbo technology ensures switchover within 10 seconds (less than one-third of the time required by its counterparts) and provides the geo-redundant DR capability.

Trusted Intelligent Computing Service (TICS):

The technique assures your security while unlocking the data value. As a key factor of production, data needs to able to flow flexibly to maximize value. However, in reality, it can still be difficult to realize openness, sharing, and circulation.

Huawei Cloud TICS delivers security for innovative and converged applications of data elements. At the same time, it enables secure data flow and maximizing data value. We can define it in the following three characteristics-

1. Based on multi-party federated data exploration and modeling, it processes privacy data in place, keeping data usable but invisible.
2. A federated AI algorithm and a homomorphic encryption algorithm are used for collaborative optimization. For example, batch computing of encrypted data is supported during model training, improving the training performance tenfold without the need for an accelerator card.
3. It supports interconnection with some of the most used data sources, adapting to typical big data application scenarios without conversion.

Pangu Model:

This includes the world’s largest Chinese NLP model and vision pre-trained model. As the account of the fact that AI has become a new driver of productivity. However, the development of the AI industry is still in the initial stage.

Based on the “pre-training + downstream fine-tuning” model of the AI industry, the Pangu Model is accelerating the introduction of AI into a wide range of industries and is making inclusive AI a reality. It has three characteristics:

1. Strong generalization ability
2.  Small-sample learning
3. High model accuracy

A single model can be applied to a large number of complex industry scenarios and can achieve high accuracy even with a small number of samples. While addressing the needs of scenarios, these models include the flour flavors-

1. Natural language processing (NLP)
2. Computer vision (CV)
3. Multi-modal models
4. Scientific computing

It’s said to be the first Chinese NLP model with hundreds of billions of parameters. At the same time, the Pangu CV model is the industry’s largest CV model.

Infrastructure software:

This software infrastructure enables diverse computing. As cloud innovation depends on the support of infrastructure software root technologies such as operating systems, databases, and middleware.

Huawei has been created a complete set of basic computing software stacks, including the operating system, enterprise-grade database, and all-scenario AI computing framework, to enable diversified computing power and support Huawei Cloud innovation.

As of now, more than 2.4 million software developers have approached the Huawei Cloud Kunpeng. Furthermore, we have previously reported that Huawei Cloud IaaS ranks No. 2 in China and No. 5 in the world, making it the fastest-growing mainstream cloud vendor.

To date, Huawei has worked with 72 universities in China, offering more than 1,000 courses, training more than 1,500 teachers, and the company plans to provide software development training to 3 million university developers over the next five years.