2022 Feitian Technology Summit: How Hard City Builds a Digitally Intelligent Electronic Industrial Internet Platform Based on SAE

The era of global digitalization has arrived, and the digital economy is driving profound changes in production methods, lifestyles, and governance methods, and has become a key force in reorganizing global factor resources, reshaping the economic structure, and changing the global competition pattern.

Cloud is a technology platform that connects reality and virtual twin worlds. It has a broad space for imagination. As the technical direction of the digital economy, cloud-native technology is becoming a joint part of the rapid development of enterprises at an unprecedented speed. On August 11, the 2022 Feitian Technology Summit was successfully held in Shenzhen. Hard City, as a member company of Alibaba Cloud Cloud Native Accelerator, was invited to attend the conference, and CEO Li Liuqi delivered a keynote speech: Hard City Cloud Native Transformation Road.

This article is based on the sharing of Mr. Li Liuqi, CEO of Hard City, at the summit.

background

Hard City is an industrial Internet platform for the electronics industry, providing a BOM-based one-stop for 100,000 enterprises including new energy vehicles, industrial control, medical care, IoT, etc. from hardware design selection, chip supply, to manufacturing, marketing and service. delivery. After these hardware facilities are designed, Hard City can provide full life cycle delivery of PCBA projects. In the process, Hard City is reducing the 2 to 3 months between product design and output to 2 weeks. In order to improve efficiency, a lot of data and technology needs to be done. For example:

1. Hard City has built the industry's leading product database, commodity database, and knowledge base for corresponding chips. This is to help Hard City reduce the one-month time that previously required manual processing by 90% when dealing with program design. , which is a few seconds plus ten minutes for manual review.

2. In the production process, it is necessary to digitize the entire production process, quality and after-sales service of the factory, so Hard City can provide the most basic products, production and service infrastructure for 100,000 hardware companies.

Problems faced by traditional architecture

In this process, we will encounter many data and technical problems. For example, the traditional technical architecture encounters three core problems:

The first is the uneven utilization of server resources . When in the peak and trough of resources, due to the uneven load, individual servers will be under higher pressure, the response will be slow, and the resource utilization efficiency will be lower.

The second is the high cost of operation and maintenance . In the previous ECS-based architecture, a shell script had to be written every time a release was upgraded, and a release was only made when the number of customers was relatively low. The release efficiency is low, and the rollback is also very troublesome; when there is a problem, the troubleshooting efficiency is low, and there is a lack of application-level monitoring.

Third, the assignment of operation authority is cumbersome . Developers and operation and maintenance personnel need to log in to the system from time to time for maintenance, and it is difficult to converge the permissions and the risk is high.

Why choose SAE

Based on the problems mentioned above, we have the following solutions:

1. Optimize our script. It is to use traditional methods to improve efficiency, but this can only alleviate the problem, not the fundamental problem, and the maintenance cost is high.

2. Solve the problem through K8s containerization. But this is more technically demanding and has a certain technical steepness. While Hard City has strong technical capabilities, we don't think this is the best way to go.

3. Select SAE. After coming into contact with Alibaba Cloud's cloud native technology, we finally chose to graft our K8s on the basis of SAE. After deploying new scenarios and applications on SAE, we found that SAE has brought us great improvements in many aspects. For example, when the original architecture remains unchanged, we can use the 0-code transformation method, without any action, and one-click migration. The second is to build a pipeline based on cloud effects, which can be very simple to complete automatic release. The third is that you don't need to log in to the machine, but you can assign permissions to development and operation and maintenance based on the permission assistant to further improve security. Of course, we also mentioned that SAE's extremely flexible capabilities can make our resource efficiency more efficient.

SAE usage practices

You can see the actual situation through the figure below.

In the leftmost picture, we can configure elastic policies for each application based on elastic rules, and can observe the elastic scaling of resources of each application. In the middle picture, you can see the application monitoring situation of each application, such as API interface call delay, slow SQL query, etc. On the far right is to focus on TOP applications, such as which application consumes the most resources and has the most abnormal requests. As an operation and maintenance, you can visualize the health status of each application in real time. These capabilities greatly reduce the burden of operation and maintenance. Especially, SAE also supports functions such as one-click rollback and efficient allocation of permissions.

Hard City is all about being more focused on the business and bringing significant changes to the business. For example, after using SAE , we have improved development efficiency by 50%. In terms of operation and maintenance efficiency, it has increased by 70%. The cost of the new application of the entire technology has been reduced by 40%, and the efficiency of the entire troubleshooting has been increased by 100%.

Specifically, as the CEO, my obvious feeling is that every time a version was released, it was necessary to avoid the peak period of users. For example, it is necessary to choose to release the version in the middle of the night, and after the release, it will bring many uncertain problems; now based on cloud-native products, users are not aware of the release, and the release time is greatly shortened, and we can even do some gray work. degree configuration.

The second is free operation and maintenance. In the past, we had a lot of systems and operation and maintenance, and we needed 2 people for operation and maintenance. Now our operation and maintenance is 0.5 people, the operation and maintenance efficiency is greatly improved, and the operation and maintenance personnel can focus on the business itself.

The third is extreme flexibility. Each new application, such as DFM and DMA in the manufacturing field, as well as the analog production of link design and complete PCBA, requires very strong computing and data capabilities, and the demand for resources is different. For example, The peak period is 10 times or even 50 times the low peak and valley. How to allocate resources efficiently and make the system more flexible? With SAE, there is now no need for human intervention, but full automation.

It is more convenient in monitoring and makes troubleshooting more visualized. Cloud native makes the entire operation and maintenance, resource allocation, and management one-click and fool-like, greatly reducing the cost of hard city. In the future, I look forward to working with Alibaba Cloud to help more enterprises to digitally innovate and upgrade.