中兴
Baidao helps ZTE with business deployment
Customer background
ZTE Corporation is a global leading provider of comprehensive communication solutions and the largest listed communication equipment company in China. Its main products include 2G/3G/4G/5G wireless base stations and core networks, IMS, fixed network access and bearer, optical networks, chips, high-end routers, intelligent switches, enterprise networks, big data, cloud computing, data centers, mobile phones and home terminals, smart cities, ICT services, as well as signal transmission equipment for aviation, railways, and urban rail transit.
ZTE provides innovative technologies and product solutions to telecom operators in more than 160 countries and regions around the world. Through a full range of wireless, wired, service, terminal products and professional communication services, it meets the differentiated needs of different operators globally.
ZTE provides innovative technologies and product solutions to top operators in more than 180 countries and regions around the world (such as China Mobile, Vodafone in the US, Telekom in Germany, Telefónica in Spain, etc.). Through a full range of wireless, wired, business, terminal products and professional communication services, it meets the differentiated needs of different operators globally.
Business requirements and challenges
ZTE’s business has spread to more than 160 countries around the world. The IT infrastructure within the group is extremely large, and the processes are extremely redundant and complex. The launch of a version requires multiple layers of approvals from various parties, and the cycle is basically 1 – 2 weeks. The internal IT toolchain is also
- The 1-hour deployment time forces ZTE to invest a large amount of engineering time in the deployment process.
- Fixing a single error requires the attention of the entire engineering team, thus delaying the release of new features.
- It is not conducive to agile development – today’s services are updated almost daily, and a monolithic architecture makes this time – consuming.
- Low reliability – The monolithic architecture builds all modules in a single process. When there is a bug in one link, it is easy to affect the entire system.
Solution
In the project, Baidao, a Google partner, helped ZTE’s IT team conduct a detailed analysis of the cost investment plan and optimization solution for the cloud – based business on Google Cloud. At the same time, it provided very good answers and business solutions to various problems and requirements. During the transformation process, GKE was used to further optimize the serverless architecture. Its professional and meticulous service support left a deep impression on the client.
The architecture design diagram is as follows:
GKE helps optimize costs and experiences
As a fully compatible K8S service launched by the GCP platform, GKE implements the cloud – native concept, fully docks with K8S in all functions, and is completely compatible with the K8S ecosystem. As a promoter of K8S open – source and a member of CNCF, Google gives GKE unique technical advantages. The newly launched Autopilot mode, from the user side, reduces users’ operation and maintenance costs; from the product side, it enables users to focus more on business pods without incurring additional costs. PODs can undertake more business traffic through automatic scaling, achieving a better combination of cost and business.
Combining GKE with HPA (Horizontal Pod Autoscaling) and VPA (Vertical Pod Autoscaling) has significantly reduced costs. During the business idle period, it can quickly save approximately 50% of the costs. During the business peak period, it can rapidly expand nodes and pods within minutes to handle sudden traffic surges. All of this is imperceptible and transparent to users, greatly enhancing the user experience.
CI and CD create a one-stop devops development model
By designing a one-stop continuous integration and continuous deployment process, the time for development and deployment to production has been saved by 80%.
Let the concept of agile development run through the entire code development process.
Among them, Jenkins, as the main tool for source code compilation and building, cooperates with Dockerfile to create an image, upload it to the Artifact Registry private repository on Google Cloud, and the GKE platform directly pulls the image and releases it to the container. Compared with the previous state of separation between development and operation and maintenance, the currently designed CI/CD system can quickly carry out operations such as system launch, deployment, and rollback.
Use the product
- Compute Engine
- Google Kubernetes Engine
- Load Balancer
- Artifact Registry
- Cloud CDN
- Google Cloud Storage
- Cloud functions
Customer earnings
Note: [All the following statistical data are from the statistical results of the customer background]
- Leverage the auto – scaling feature on Google Kubernetes Engine to improve availability during peak loads, reduce infrastructure hosting costs by 40%, and further cut down on operational labor costs. Due to the use of the managed mode, customers don’t need to worry about business interruptions caused by cluster version updates or node failures. They can spend more time on architecture optimization and operations. From the previous situation where three people were needed to maintain the self – built K8S, now only 1 – 2 people are required for maintenance. This results in a 33% – 67% savings in labor costs.
- Improve the network connection stability by 20% through Google’s high-quality dedicated network.
- Scale on demand according to the business development speed, achieving zero latency in hardware procurement, configuration, and maintenance.
- As a professional big data analysis tool of Google Cloud, Dataproc has reduced the customer’s maintenance time by 60% compared with the Hadoop cluster previously built by the customer. On the operation side, it has also saved a large amount of data analysis time, enabling the company’s strategies and marketing activities to be implemented more quickly.
