Cloud native provides more possibilities for developers

Bbackground

Cloud native is an inevitable product of the development of cloud computing, and the continuous growth of cloud native is by no means accidental.

What will cloud native look like in 2021, and what new changes will it bring? Yi Li, R&D Director of Alibaba Cloud Container Service, recently delivered a speech "The New Boundary of Cloud Native Applications" at the Alibaba Cloud Developer Conference, and said that cloud native provides developers with three conveniences: "zero" maintenance of application infrastructure, application Architecture modernization "zero" resistance, digital and physical world "zero" boundary.

Cloud Native: Born for the Cloud

Cloud native is a technology born of the cloud. It is rooted in developers and provides maximum cloud value.

In the CNCF 2020 Developer Status Report, more than 4.7 million developers around the world are now using cloud-native technologies, accounting for 36% of all back-end developers. Developers have become the most important driving force for cloud native transformation.

Application infrastructure "zero" maintenance

Cloud-native technologies such as containers and serverless continue to push the computing interface up and the complexity down, allowing developers to focus on business innovation instead of infrastructure, which can greatly improve R&D efficiency.

Alibaba Cloud provides developers with the most abundant cloud-native products in the country, helping enterprises to focus on business innovation rather than infrastructure construction. Enterprises can realize Internet-based application architecture through container services, function computing, and service grids. On top of this, products such as cloud-native databases, cloud-native AI, and cloud-native big data can help enterprises accelerate the digitization and intelligence of business processes. change.

"Zero" resistance to application architecture modernization

More and more enterprises hope to bring new benefits through application modernization, such as micro-service and Mesh, and better meet the needs of business development. However, new technologies will also have a great impact on existing application architectures. Using cloud native technology, the existing application architecture can be smoothly upgraded step by step.

When modernizing existing applications, developers need to split a single application into a distributed microservice architecture. Microservice architectures such as Spring Cloud / Dubbo build service governance logic in the form of SDK code bases in the application. But there are several problems with this architecture:

Intrusion: In the microservice framework, the implementation and lifecycle of service governance capabilities are coupled with business logic. Changes and enhancements to service governance capabilities require rebuilding and deployment of applications, resulting in higher upgrade and maintenance costs.

Implementation binding: Since the microservice framework code base is usually implemented by a specific language, it is difficult to support the integration between heterogeneous systems in multiple languages (polyglot) as a challenge.

Therefore, the community proposes the Service Mesh (Service Mesh) architecture—decoupling application business logic from service governance capabilities. The capability of service governance runs in an independent sidecar process and is deployed independently. Through network interception, capabilities such as service discovery, traffic management, observability, and security that are transparent to applications are realized.

Solve the above intrusive and binding problems, the specific advantages are as follows:

*Sinking of complexity: service governance is implemented to sink to the infrastructure, which can evolve independently. This allows developers to focus more on the business application itself.
* Zero intrusion: High-level capabilities such as zero-trust security and observability can be realized without code modification.
*Multi-language support: It can transparently support multiple programming languages and programming frameworks.

So, are microservices and service grids either one or the other, and you can’t have both? How to be compatible with the existing microservice architecture while transforming the service grid?

With the efforts of the community, service mesh and microservices can be well combined to support the smooth evolution of enterprise microservice architecture.

Taking Dongfeng Nissan Motor as an example, this paper introduces the service grid migration path of the enterprise. First, its data services are developed in different languages such as Python/Java. Java applications use the Dubbo microservice framework, and Python uses REST/HTTP for service calls, lacking unified service governance capabilities; Ways coexist, hoping to fully migrate to the container architecture.

Through the ASM service grid, no matter whether it is a Python/Java application, a virtual machine or a containerized deployment, it can join the service grid to achieve service governance in a unified and declarative manner. Among them, the existing Dubbo microservice applications and applications in the grid can use the Nacos registration center to realize service registration and discovery, and maintain the compatibility of the existing application architecture.

Digital and physical world "zero" boundary

Digital innovation needs to go deep into the industry and integrate the physical and digital worlds to realize innovative business value. The integration of cloud, edge and terminal computing has become a trend. Yesterday's Alibaba Cloud Summit described the direction of cloud development in the future. One cloud has multiple cores, one cloud has multiple forms, and the combination of cloud and AIoT can support ubiquitous computing. The cloud-native technology represented by containers will become the most important carrier of the next generation of distributed cloud applications because of its advantages of agility, light weight, and portability.

Logistics is a typical scenario of digital innovation. It revolves around the four dimensions of people, goods, machines, and vehicles, and involves complex business scenarios such as massive data processing and intelligent scheduling. Taking STO Express as an example, it involves the transfer, transportation and delivery of hundreds of millions of packages every day. Digital technology plays an important role in the optimization of logistics supply chain. STO Express has built a logistics cloud PaaS platform with an overall cloud-edge-end integrated architecture based on Alibaba Cloud edge container products.

The PaaS platform is responsible for distributed resource scheduling and application management, big data processing and intelligent analysis in the central cloud.
The edge cloud nodes located in various storage centers combine IoT devices to support the core process of the express delivery business. Operations such as scanning and verification can be completed locally, reducing delays and strong dependence on the cloud.
This architecture can help enterprises reduce costs by 30%, and increase stability from 99.9% to 99.95%. It not only supports daily business development, but also calmly handles business peaks such as Double Eleven.

Cainiao Logistics Cloud PaaS uses Alibaba Cloud's edge container service ACK@Edge to solve the challenges of distributed resource scheduling, application management, and autonomous operation and maintenance after computing sinks. The core technology behind it is Alibaba Cloud's open source OpenYurt project, which has become a CNCF sandbox project.

Edge computing faces technical challenges such as distributed computing power, heterogeneous resources, and weak network connections. OpenYurt is a cloud-edge collaborative computing framework based on Kubernetes. It has core capabilities such as edge application management, edge autonomous self-healing, and edge computing power management.

In addition, OpenYurt insists on non-intrusive implementation in native K8s, focusing on standardization and openness. In the past two years, OpenYurt has been implemented in industries such as CDN, Youku, Cainiao, Industrial Brain, and City Brain, and has also supported customers such as Acoustics and Kuaishou.


If the cloud is the intelligent brain of an enterprise, IoT devices are the eyes and hands, realizing the interaction with the physical world. Using K8s to reduce the management complexity of massive distributed devices can achieve unified management and better collaboration of distributed applications and IoT devices. Combining cloud-native and IoT creates enormous opportunities for innovation.

Join hands with VMware to build a cloud-native IoT ecosystem, and gather open source communities to create industry standards
Yi Li, Director of Alibaba Cloud Container Services, and Lu Guang, R&D Director of VMware China R&D Center, jointly announced that they have reached a technical strategic cooperation in the field of "cloud-native edge computing". Jointly promote the process of cloud-edge integration and help more enterprises fully embrace the transformation and upgrading of digital intelligence.


Based on common ideals and visions, the OpenYurt community and the EdgeX Foundry community under the Linux Foundation will cooperate in depth in the fields of edge computing, IoT, and cloud native: on the one hand, redefine the device management model in the IoT field through cloud native methods, and realize device twinning capabilities ; On the one hand, and use EdgeX Foundry's mature technology ecosystem to allow cloud-native applications to support various IoT protocols and devices.

The in-depth cooperation between OpenYurt, an open source project of Alibaba Cloud, and EdgeX Foundry, an open source project jointly initiated by VMware and maintained by its Chinese community, will help enterprises and edge business developers to easily build cloud-edge-end integration without any modification of K8s Coordinated IT architecture. As a "plug and play" open source IoT Edge platform, EdgeX Foundry (EdgeX) supports devices from different manufacturers using different protocols. At the same time, OpenYurt can extend Kubernetes to edge scenarios through native plug-ins, and supports all upstream Kubernetes features.