"Cloud" on the code - "Cloud Native. Cost Reduction and Efficiency Increase" e-book review

After being recommended by a friend, I participated in the activity of "Motive Force. Cloud Native. Cost Reduction and Efficiency Increase". Cloud native is an emerging software development and deployment methodology, which aims to take advantage of cloud computing technology to achieve more efficiency, flexibility and reliability. application development and deployment.

1. Benefits of reading e-books:

If you are not interested in reading e-books, you can watch the video lecture directly: https://marketing.csdn.net/p/6c2a12739080d8fba0fb0b529a656de1

"Cloud Native. Cost Reduction and Efficiency Increase E-Book" is an e-book jointly written by Tencent Cloud, China Academy of Information and Communications Technology, FinOps and CSDN. A large number of examples and cases give quantitative comparative data.

By summary, it is roughly divided into the following parts:


1. The value of enterprise cloud-native applications

insert image description here

Cloud native technology has practical applications in various industries, and introduces how to solve the challenges encountered in the implementation process, so as to improve the resource utilization rate of enterprises and realize the reduction of operation cost and efficiency increase.
insert image description here


2. Cloud native cost reduction and efficiency increase practice plan

insert image description here
insert image description here

The e-book provides a large number of design patterns, technical principles, strategies and best practices of Tencent Cloud in cloud-native applications :

. 结合这些策略和实践不仅能成为一种标准化work instruction作业流程,更能够帮助我们更系统、原理性的了解云原生技术。

②. 让我们能够得到启发和借鉴,从而良性的在企业落地和实践中规避常见的错误和问题。

③. 建立一套更高效、更安全、更灵活、高品质、可扩展、高可用、易维护的云原生技术体系。

3. Cloud native cost reduction and efficiency increase case

(1). Homework Help:

business pain point Improvements Quantitative results

1. Scale: Thousands of application services, corresponding to tens of thousands of service instances, running on hundreds of thousands of computing cores
2. Complexity: The technology stack covers multiple mainstream languages

Through cloud-native transformation, use infrastructure to take over a large number of non-functional logic in the business to achieve elasticity, observability, resilience, automation, and sustainability.

1. Using the Go language to refactor the application module can bring about a performance improvement of more than five times
. 2. Based on the cloud-native transformation of the job help, the overall cost reduction service has reached 40%

insert image description here

(2). Sanqi Mutual Entertainment:

business pain point Improvements Quantitative results

1. How to expand the capacity of multiple computer rooms in IDC operation and maintenance
2. In the IDC environment, how to improve resource utilization and stability while resource isolation?

1. From IDC to the cloud, replace the virtual machine with a cloud virtual machine architecture, and transform to cloud-native based on the deployment architecture on the cloud.
2. Such as: https certificate uninstallation, WAF interception of illegal products, etc.

The total capacity is always 5~7 times of the actual business peak . Controllable capacity means controllable cost, which also provides a good decision-making basis for cost reduction or increment, so the cost can be reduced by reducing the total capacity.

insert image description here
insert image description here

(3). Jingdong:

Business pain points:

. v1物理机部署资源利用率低、应用资源无法隔离、应用无隔离等问题

②. v2服务器资源采购受限、资源利用率有待提升、资源使用不均衡等问题

③. v3应用上下线成本高、在线集群非高峰段资源利用率不高等

Improvements:

insert image description here

(4). Tencent:

business pain point Improvements

1. Application resource usage settings are unreasonable
2. Pods of the same type have different resources
3. Multi-dimensional idle resources are severely fragmented
4. Sudden traffic peaks lead to insufficient resources and limited resource dimensions

1. Practical experience with Eunomia
2. Solvers, schedulers
3. Predictive models

Quantified results:
insert image description here

(5). Summary:

. 从上可以看到,腾讯云原生应用在比较大规模下的实际案例,充分的展示了云原生技术在企业应用中的巨大潜力和优势,体现降本增效实践的价值。

②. 文中不仅仅只是提出的思路和方向,还给出了实际帮助公司解决了成本的业务痛点(并非纸上谈兵),可以极大的推广以及应用到企业真实项目。

③. 同时,通过量化的数据和解决方案,让云原生应用程序确实可以帮助企业降低开发和运维成本,提高团队的工作效率和应用程序的可靠性,从而实现更高效的业务运营和更好的用户体验。

2. Combined with the company's business needs and development work, the implementation of cloud native implementation:

I am a Java full-stack developer. Currently, the company adopts the k8s cluster solution, but there are always problems in resource utilization. Therefore, I want to benefit a lot from this live broadcast and e-book learning, and understand my own shortcomings and defects more clearly. I hope that through learning, I can optimize the procurement cost for the company and achieve cost reduction and efficiency increase.

Consider through several directions in the above figure:
insert image description here

1. Fine-grained:

. 微服务架构可以将容器化技术拆分成多个小的、独立的服务,从而实现架构的可移植性和可伸缩性。

②. 公司也是搭建了多个容器仓库组,提供多个镜像可以快速的应用到生产中。

③. 提供了多个主机组,可以使不同的项目做到隔离。

insert image description here

insert image description here
insert image description here
insert image description here

2. Extreme elasticity:

. 目前公司的k8s遇到的困境是经常因为资源不足而被迫升级配置、或都临时下掉没有使用率高的配置。

②. 并且生产、测试、预发布经常遇到内存不足的内题,可以对比CPU的使用率还是较小,内存的使用率偏高。

③. 如果导入了动态扩容机器,感觉可以解决这部分问题点。

insert image description here
insert image description here

3. Deliver standardized CI/CD:

(1). Function description:

.CI/CD等流程加入到云原生部署流程中,能够快速进行测试、构建和交付软件

②. 通过插件化的自定义配置策略,完全可以覆盖99%的场景需求。

③. 同时,提供版本回滚、机器人集成等多种功能,节约了大量的人工操作成本,缩短上线时间。

④. 可以更快地交付产品迭代并快速满足客户需求,提高产品开发的敏捷开发迭代效率。

⑤. 提供了部署yaml脚本,可以进行灵活的配置。实现env切换。

insert image description here
insert image description here
insert image description here
insert image description here

(2). Scenario example:

. 比如将前端的脚手架打包生成一个镜像仓库,每次打包只需要下载并执行镜像中的脚手架即可打包代码

②. 大大的缩减的之前npm install组件包的时间。

(3). Later planning:

. 将sonarQube平台、灰度测试集成到自动化运维、持续交付中,可以保证团队代码的规范和最小化潜在的风险。

②. 有利于构建一个更稳定、更可靠的产品交付。

4. Unified management of asynchronous resources:

. 云原生应用中也自带了很多组件,可以很快速的构建微服务架构、容器编排、服务发现和治理、监控和告警一套服务体系。

②. 由微服务、容器化技术、服务网格技术、自动化可以让云原生技术能够动态地进行伸缩和升级。

③. 同时结合xx-job可以做到有效的扩展。

insert image description here
insert image description here
insert image description here


Fourth, crane understands:

After reading the e-book, I found a sentence that is very touching: "Improving application performance includes improving resource utilization, lowering technical thresholds, saving R&D and testing costs, and reducing operation and maintenance investment."

  1. Learn about FinOps and Crane, github: https://github.com/gocrane/crane,
    add cluster:
    insert image description here
    insert image description here

View cluster information:
insert image description here

Moreover, it is still based on Grafana monitoring, which is recommended by many colleagues and companies, and has a detailed description of multiple indicators of the server:

insert image description here


5. Experience summary and suggestions:

  • This e-book introduces the status quo, trends, best practices, and the road to enterprise practice of cloud native technology in a simple and simple way
  • From the analysis of the current situation and the challenges faced, to the explanation of various practical solutions, and finally to the practice of the company's real scene
  • By explaining the knowledge points related to cloud native technology in detail, covering multiple latitudes, and explaining cloud native technology in an all-round way, it can better help readers understand the cloud native technology system and its application value in essence, and better To effectively deal with the current and future challenges of the cloud native technology system.

Summarize:

The advantage of cloud-native applications is that they can be deployed and expanded more quickly and flexibly, reducing the waste of hardware and software resources, and can also run more stably and reliably, reducing the risk of system failure. In addition, cloud-native applications can also provide better observability and debugging capabilities, enabling development and operation and maintenance teams to find and solve problems faster, thereby improving the productivity of the entire team.

For those who are interested, if you want to know more details, you can directly read the e-book or the supporting video lecture.

Guess you like

Origin blog.csdn.net/wanmeijuhao/article/details/129992221