Summary: DevOps capabilities reflect the ability of technology development to respond to business changes. As organizations grow in size and business complexity, DevOps capabilities will become more important. Continuing to improve DevOps capabilities is a common challenge for technology development.

Editor’s note: this article from ali greetings effect produced by the team, the alibaba conversation practice guide, to: developer.aliyun.com/topic/devop…

DevOps capabilities reflect the ability of technology development to respond to business changes. As organizations grow in size and business complexity, DevOps capabilities will become more important. Continuing to improve DevOps capabilities is a common challenge for technology development.

To provide direction and a clear path for the organization’s DevOps capabilities. We defined a DevOps capability Maturity model that provides two values: 1) knowing where we are today; 2) How to plan the path of improvement.

We divide DevOps implementation into 4 categories and 10 capabilities, which are as follows:

Basic capability: includes the servitization level of the system and the infrastructure level, which is the foundation of r&d and delivery. Among them, the level of servitization is closely related to the application architecture. The most ideal case is no servitization architecture, and the lower case is the monolith architecture coupled with the whole system. The level of infrastructure is reflected in the attention that R&D needs to pay to infrastructure. The more attention it needs, the higher the cost of r&d on infrastructure, the lower the efficiency, and the stability is difficult to guarantee.

Deliverability: including tool level, test automation level and deployment and release level, it is the main embodiment of engineering capability level. Among them, instrumental level refers to the overall level of all kinds of tools involved in the whole process of R&D, including single point ability (such as project collaboration tools, construction tools, dependency management tools, environmental management tools, etc.) and collaborative ability (such as the integration of requirements and released systems, defects and tests, etc.). Test automation level refers to the feedback efficiency and automation degree of test. Test automation is an important part of engineering ability, and also the basis of improving deployment and release ability. Deployment release level refers to the ability to bring an artifact online and provide a service to the production environment, including the degree of automation, stability (such as smooth grayscale release), and adaptability (i.e., the ability to handle different situations and self-healing when problems occur). Good delivery should be consistent, fast, high quality and low risk.

Operation and maintenance capability: includes the observability of the system, the operation and maintenance level of applications, and the operation and maintenance level of infrastructure. It reflects the flexibility and toughness of the system during operation. Among them, observability is the most important part of operation and maintenance capability, mainly reflected in whether the global operation situation and problems can be seen from the perspective of the system, usually reflected in the monitoring level, link analysis and problem location ability. Application O&M refers to the operation and maintenance (O&M) of applications, including modifying configuration items, adjusting running parameters of applications, and adjusting applications, such as capacity expansion or reduction. Infrastructure OPERATION and maintenance (O&M) refers to the operation and maintenance of the infrastructure of the system, including VMS, container platforms, and basic services (such as domain names and configuration centers). The best operation is self-operation.

Synergy: including the synergy between business and technology, as well as the synergy between development and operation and maintenance, which reflects the overall synergy and business responsiveness. Among them, the coordination of development and operation and maintenance is to make the delivery process smoother and more efficient, to improve the response speed of technology, while ensuring the flexibility and toughness of system operation; Technology and business synergy aims to make the transfer and delivery of value from business to technology more precise and efficient, and feedback more immediate, to improve the efficiency and effectiveness of business development and innovation.

Maturity model

Based on these four categories and 10 competencies, we present a maturity model with five levels, from L0 to L4 maturity.

classification

Ability to

L0

L1

L2

L3

L4

Basic ability

Service level

Low (singleton or just started servitization)

(Business development based on a specific framework)

(Business development based on a specific framework)

High (business development only, language related)

High (business development only, language independent)

Infrastructure focus

High (non-cloud native infrastructure)

High (non-cloud native infrastructure)

Medium (Cloud native Infrastructure)

Lower (primary serverless)

Low (full serverless)

Ability to deliver

Instrumental level

Low (no DevOps tool chain)

Middle (partial, isolated island type tool)

High (continuous delivery tool chain)

High (continuous delivery tool chain)

High (end-to-end DevOPS tool chain)

Test automation level

Low (no automation, long feedback time)

Low (no automation, long feedback time)

Medium (partially automated)

Higher (mainly automated)

High (fully automated)

Deploy publish capability

Low (manual publishing)

Lower (automated tool release)

(Automated declarative publishing)

High (automatic declarative release, grayscale intervention ability)

High (automatic declarative release, automatic grayscale capability)

Operational capacity

observability

Low (only scattershot base resources are observed)

Low (scattered base resource and service interfaces observable)

Medium (Overall service capacity is observable)

High (panoramic business capability and service capability observable and traceable)

High (panoramic business capability and service capability observable and traceable)

Apply o&M capabilities

Low (Human operation and Maintenance)

Low (Automated tool operation and maintenance)

Medium (declarative operation and Maintenance)

Medium (declarative operation and Maintenance)

High (Service self-operation and maintenance)

Infrastructure operation and maintenance capability

Low (single point, human operation and maintenance)

Low (single point, automated tool operation and maintenance)

Middle (cluster, automatic tool operation and maintenance)

Middle (cluster, automatic tool operation and maintenance)

High (cluster, automatic operation and maintenance)

Coordination ability

Develop, operation and maintenance coordination ability

Low (separation of development, operation and maintenance

Batch deployment, independent operation and maintenance)

Low (development, operation and maintenance cooperation deployment, increased deployment frequency)

Development of independent deployment

Development assist application operation and maintenance)

High (the development team continues to deliver

Integration of development and operation teams)

High (continuous delivery and self-maintenance by the development team)

Ability of business and technology coordination

Low (business and technology independence, throw requirements, little synchronization)

Low (regular synchronization of business and technology, batch development, deployment and release)

(Continuous development, deployment and release based on business requirements)

(Continuous development, deployment and release based on business requirements)

High (quick response and delivery of business requirements, immediate feedback and adjustment of business requirements)

L0: Manual batch delivery, manual operation and maintenance, this is the DevOps stage with zero capability. The service capability of the DevOps completely depends on the individual developer. The quality of business delivery is generally not high.

L1: Handicraft, tools, auxiliary batch delivery and operations, this stage began operations, introducing automated tools to assist in the release, usually have the foundation of service, infrastructure has been part of the cloud, and through the introduction of open source tools or self-built built some tool to fulfill certain specific demands, but these tools are often isolated, no link, Business, development, operation and maintenance are synchronized on a regular basis, and requirements are delivered in batches.

L2: Based on the needs of the business part of the automated delivery operations, this phase is sustained release based on business requirements, and has adopted the declarative operations, usually native already use cloud infrastructure, and use the cloud resource management service status, most of the tool chain can already together, to achieve a certain degree of continuous delivery. Services begin to have middleware level abstraction and governance capabilities, but generally cannot be self-operated and maintained, and rollback operations still need to rely on human judgment and processing.

L3: End-to-end automated delivery and limited self-operation and maintenance based on business requirements, the delivery frequency and delivery quality of business requirements in this stage have been significantly improved, and the level of servitization has been quite high. In most cases, business development can be focused on specific technology stack, and the main services are published and run in serverless mode. The publishing process can be automated and declarative, only requiring a small amount of intervention in gray processing, and the service has been able to achieve self-operation and self-governance in most cases.

L4: the end-to-end continuous delivery and adjustment of business requirements is closed loop and fully self-operated. At this stage, developers only need to pay attention to business development, and business requirements can be delivered and adjusted quickly. The level of sertization is decoupled from the technology stack and fully serverless. The entire delivery process is fully automated and the service is fully self-governing. This stage is what we are looking for.

The original link

This article is the original content of Aliyun and shall not be reproduced without permission.