On May 19, 2023, hosted by the China Academy of Information and Communications Technology (hereinafter referred to as "CAICT") and the China Communications Standards Association, organized by the Cloud Computing Standards and Open Source Promotion Committee of the China Communications Standards Association, and supported by a co-construction and sharing platform for digital development of enterprises The "2023 Low Code No Code Industry Conference" was held in Beijing.
With the theme of "Reshaping the Digital Engine and Enjoying a Modern Chapter", this conference announced the latest evaluation results of "Low Code No Code" and released "2023 Low Code No Code Application Case Collection", submitted by Alibaba Cloud Computing Co., Ltd. Two cases were selected into the no-code case collection.
Automobile industry
Geely Automobile Group Co., Ltd. - Geely Automobile Intelligent Application Acceleration
Program introduction
Through the Aliyun Feitian operating system, the underlying computing power is uniformly controlled and deployed, supporting the cluster expansion of 100,000 servers in the Geely Star Computing Center, and through the optimization of storage and network communication technologies, the parallel computing efficiency of kilocalories is improved to 90. % or more, the utilization rate of computing power resources is increased by more than 3 times. Interconnect with the public cloud to maximize the advantages of the cloud.
Combined with the IDC network environment and Al large model training requirements, a two-layer HDR SW network architecture solution is adopted. The two-tier network is HDR MCS network cluster and HDR MQM cluster respectively. Each GPU server is connected to 2 MQM nodes through 4 SWCs. There is no single point and no blocking. The maximum transmission speed reaches 200GB/s, which is 20 times that of Ethernet. It only takes 8 Hours.
The advanced R&D software tool chain leads the standardization and efficiency of product R&D. According to the different business classifications of automotive product R&D and operation, it provides industry-leading big data R&D platforms, artificial intelligence R&D platforms, and intelligent driving R&D platforms (multi-platform data interoperability) software tool chains.
For multi-platform business, use the visual architecture design capability of Cloud Speed CADT to plan in advance and output a deliverable standard cloud product architecture to form a standardized specification.
Configure cloud products, regional planning, and product parameters required by the business in the architecture template by dragging and dropping. The intuitive architecture perspective helps accurate verification, improves management and use efficiency, and forms standard specifications. Through architecture replication and batch deployment, Improve operation and maintenance efficiency.
demand analysis
China's auto industry is shifting from the first half of "electrification" to the second half of "intelligence". What can support the transformation of intelligence is a digital system that can fully cover R&D, manufacturing, and sales. Geely is moving towards the goal of "building a world-class good car". The opening of Geely Star Wisdom Computing Center is a landmark event, which will help Geely to develop technologies in the fields of intelligent driving, intelligent cockpit, artificial intelligence, and new energy power management. intergenerational advantage. There are also higher requirements for the underlying basic cloud platform, which requires not only sufficient computing power support, but also standardized architecture design and stable and easy-to-use product capabilities.
- How to connect R&D, manufacturing, sales/after-sales, operations and other data, so as to achieve unified data storage, governance, analysis, and operation, and help enterprises transform to intelligence;
- It is imminent to establish a "cloud-data-intelligence" integrated data center that supports data closed-loop business, data capitalization, security compliance, and facilitates innovation in production and research;
- It is necessary to transform product design into a visual cloud architecture solution, construct according to the diagram and improve the high availability of infrastructure;
- It is necessary to carry out fine-grained management of the cloud, from project resource evaluation, cloud product quotation, fast delivery and operation and maintenance, to form a complete and efficient life cycle management.
Solution effectiveness
Geely Xingrui Computing Center is based on the cloud computing, big data and AI integration advantages of Alibaba Cloud Feitian Smart Computing Platform, combined with leading computing power scheduling management algorithms and R&D systems, which has improved Geely's overall R&D efficiency by 20%. For multi-platform applications, rich cloud resource products are required; multiple environments are deployed; the same type of business has similar characteristics to the underlying resource structure. Therefore, Geely Xingrui Computing Center has designed the same set of architecture templates, which can be reused in multiple environments and multiple projects.
- Unified data into the lake: Geely Xingrui Computing Center has realized the unified collection of business data of each section, forming a comprehensive data lake of the group, and the value of data has been greatly improved. No matter in the field of technology or industry, it is the first of its kind.
- R&D tool chain development: Geely Xingrui Computing Center has realized end-to-end full-process R&D support. It can simulate 100,000 kilometers of intelligent driving behavior training in one day. Combined with the professional proving ground and actual road test of intelligent network connection, it can meet the L3 global certification standard. It has now supported the research and development of the intelligent driving mass production business within the Geely system.
- Architecture visualization: Based on the cloud-based CADT, the whole link from architecture design, evaluation, verification to delivery and deployment is visualized, and cloud specifications are formed to flexibly plan and evaluate architectures based on graphs, and the application achieves 100% accuracy landing.
- Architecture replication: identify potential problems and reduce operation and maintenance error rates through early verification. By depositing templates, multi-environment and multi-region basic resource architecture replication and deployment are performed. Improve the efficiency of cloud product usage and shorten the delivery cycle of major projects from weeks to hours.
Application field
car, smart car
entertainment industry
Chengdu Sobey Digital Technology Co., Ltd. - Multi-region unitized service of cloud architecture resources
Program introduction
Alibaba Cloud CADT can implement one-time and distributed deployment of a large number of application servers, Kubernetes clusters, databases, storage, load balancing and other basic resources contained in multiple availability zones in cloud nodes, which satisfies the needs of news Safe, stable, and fast deployment of cloud and public cloud resources.
- Through Cloud Express CADT, the overall system can be configured in a visualized architecture diagram, and cloud resources can be dragged to the architecture diagram canvas to enable all resources with one click. Compared with the console, the efficiency of resource activation is greatly improved.
- Quickly configure the network environment through CADT, VPC, switch network segment design, cloud enterprise network configuration, global acceleration server monitoring group configuration, accelerated bandwidth package binding and accelerated regional bandwidth allocation.
- Through CADT, multiple resource types of Alibaba Cloud are integrated and configured, including the relationship between resources, and the binding relationship between resources is configured when cloud resources are activated.
- Apply resource planning through CADT to quickly plan the availability zone selection for resources with tight inventory.
- Through CADT to calculate the price based on the architecture diagram, you can conduct batch inquiry in real time, synchronize account discount information, and quickly make a price budget for regional server opening.
- Share the architecture diagram with Alibaba Cloud architects through the CADT private template. The architect can quickly adjust the architecture diagram and share it with customers. Customers can create applications based on templates to implement one-click deployment.
- Through CADT, the overall structure of the system can be checked, adjusted, and optimized at any time after the system is built.
demand analysis
Under the pattern of "Internet +" driving industrial innovation, radio and television media are ushering in an opportunity for development. The future of television will be integration of stations and networks, user participation, based on cloud computing, mobile Internet, big data, social networks, etc. A new TV medium for Internet business. The Internet has been deeply embedded in all aspects of TV development, becoming a new engine to promote media integration and development, and creating a new media development ecology. This means that there is an urgent need for a media production model centering on the "five integrations" of content, technology, platform, operation, and management to achieve a deep integration of TV production and Internet production that complement each other.
A key customer of Sobei’s long-term service urgently needs to migrate from the offline IDC to the cloud. Taking advantage of Alibaba Cloud’s deployment in multiple data centers across the country and around the world, it quickly deploys the converged media business solution to Alibaba Cloud’s six different regions at home and abroad. And quickly complete the unitized deployment of the business:
- The design of the network segment is complex, and the network segments cannot overlap when deployed in multiple regional availability zones.
- The configuration of Alibaba Cloud infrastructure is cumbersome, and the code development of the architecture is time-consuming. Each region has different availability zones and resource inventories, and the deployment code cannot be reused.
- The deployment of Alibaba Cloud resources covers as many as 17 types of IaaS and Paas.
- Cost accounting is difficult, and there are many types and specifications of resources.
Solution effectiveness
The characteristic of the converged media business system is that there are many types of cloud resources, which have the characteristics of the Internet. They need to enter and exit the network, and they need to serve customers nearby. Therefore, they need the same architecture to deploy services in multiple regions. Each region also needs to be able to connect quickly. .
- Batch delivery of resources: Unified configuration of the overall architecture diagram, one-click activation of all resources, and improved efficiency.
- Quick copy of cross-district city architecture diagram: quickly copy the specifications and attribute configuration of each basic resource, cloud database automatically adds whitelist, network load balancing automatically binds public network IP, cloud server automatically binds virtual server group, and cloud resources are bound in batches Resource groups, cloud resource batch maintenance tags, security group port configuration, designed once for a single region, quickly copied to other regions.
- Quick configuration of multi-regional network connectivity: quickly build a cloud enterprise network through the architecture diagram connection, global acceleration.
- Business ownership governance: batch label resources and assign resource groups through the global architecture diagram.
- Cloud-based sharing of private templates: improve the efficiency of remote discussion of architecture diagrams, and truly realize the working style of architects on the cloud.
During the implementation process, Alibaba Cloud CADT took full advantage of its own unit architecture design and multi-regional advantages, shortening the deployment time of resources required by each cloud node from the estimated 2-3 days to less than 30 minutes, greatly improving the efficiency of resource deployment and activation , to ensure the smooth operation of the project.
Application field
Cloud service, multi-region business deployment