Cloud computing moves towards serverless

On November 3, at the 2022 Hangzhou Yunqi Conference, Zhang Jianfeng, president of Alibaba Cloud Intelligence, said that a new computing system with the cloud as the core is taking shape, and the software research and development paradigm is undergoing new changes. Serverless is one of the most important trends. Alibaba Cloud will firmly promote the comprehensive serverless of its core products to help customers better realize agile innovation.

This article wants to discuss with you the inevitability of cloud computing moving towards Serverless.

Evolution history of software architecture and R&D model

The business development appeal of enterprises has always been the main force to promote the evolution of software architecture and R&D mode. Enterprises always expect to be able to respond more quickly to the growth of business scale and complexity, bring products to the market faster, and speed up business innovation, which requires technology to achieve rapid iteration of large-scale and complex software.

The traditional enterprise application architecture is usually single, and all modules are coupled together and released at the same time. This single architecture application is easy to manage at the beginning, but with the development of business, it will bring about a significant increase in complexity. This strongly coupled architecture brings about a lot of conflicts in the development, testing and operation and maintenance processes, which slows down the entire iteration.

For example, the development of the entire application requires that all modules use a unified language and framework technology stack. If a basic library is shared by multiple modules, and one module wants to upgrade to a new version, you need to persuade everyone to upgrade at the same time, even if others do not need a new version. The release rhythm of all modules is forced to align, and the issue of one module will affect the release of the entire application. It is also very difficult to quickly repair the online problem of a module, because it requires merging with the changes in progress of other modules, resolving conflicts, republishing the entire application, and running all tests before it can be republished online.

So soon, the single application architecture could not meet the requirements of software R&D efficiency, and was replaced by the Internet distributed architecture with micro services as the main feature.

After adopting the microservice architecture, the application is composed of independent services. These services are loosely coupled and interact through API calls, event triggers or data flows. Each service completes a specific function and is independently developed, run, and published. Microservice solves the bottleneck of R&D efficiency of single architecture, but it puts forward very high requirements for application infrastructure. For example, to ensure that independently developed microservices can coordinate as expected, detailed integration and end-to-end testing are required. The number of application deployments in the test environment is usually 10 times that in the production environment. If the application infrastructure cannot quickly provide an independent testing environment, a large amount of testing time will be spent on solving the problem of environment stability.

According to the research and development statistics of Alibaba Group, the research and development of one man day usually corresponds to the test of 5-7 man days. The test environment has become the biggest pain point for Alibaba Group to improve its R&D efficiency. The loose coupling of microservices also brings great challenges to database usage, state management, problem diagnosis, and application delivery pipeline. There has been a lot of discussion about the complexity and solutions of microservices in the industry, so I won't repeat them here. It is the consensus of the industry that the Internet distributed architecture with microservices as its core is highly complex to implement and must be supported by good tools and platforms.

In the software R&D process, all work can be divided into the following three categories:



1. Business code development to realize business logic.

2. Non functional code development, including the ability to achieve fault tolerance, security, observability, operation and maintenance, and third-party software integration, has nothing to do with business logic, but is a must for enterprise applications.

3. Application infrastructure management. Including building development, testing, production environment, resource planning, security control, etc.

Of the three types of work, only the first type brings real value to the business and is closely related to the core competitiveness of the enterprise. However, with the increase of software complexity, a large amount of R&D resources have been consumed for Type 2 and Type 3 work. It is the inevitable direction of software architecture and R&D mode development to reduce the complexity of Category 2 and Category 3 work as much as possible and let customers focus on business logic development.

In the past decade, both open source communities and cloud vendors have abstracted non functional code development and application infrastructure management into standardized and reusable software/services in different fields.

Cloud computing moves towards serverless

Serverless is a very broad concept, not limited to computing. Generally speaking, a service that meets the following conditions can be called a Serverless service:

1. Full custody service. It means that customers use abstract service-oriented interfaces instead of directly facing the underlying resources, so there is no burden of installing, configuring, maintaining or updating software and hardware. Full custody services usually also provide built-in fault tolerance, security and observability capabilities, which users usually do not need to rebuild.

2. Adaptive elasticity. It means that the service can automatically and elastically scale according to the load size, which greatly improves the resource utilization efficiency.

3. Pay according to the actual usage. This means that you only need to pay according to the actual execution time, traffic or number of calls, reducing the cost.

Therefore, the core value of Serverless service is to eliminate the non functional code development of customers as much as possible, simplify the application infrastructure management, and achieve a leap in R&D efficiency.

Zhang Jianfeng, president of Alibaba Cloud Intelligence, said that Serverless has turned cloud computing from a resource to a capability. In the past, cloud computing has replaced physical servers with cloud servers, but customers still purchase cloud resources in the mode of "several cores and several gigabytes of servers". In the future, cloud computing will be serverless, closer to the "grid" mode, and pay according to the number of calls.

As the Serverless service fits in with the trend of the evolution of the R&D model, we can see that in the past ten years, various cloud manufacturers are constantly turning their product systems into Serverless, and most of the new cloud services or functions are in the form of Serverless. A large number of open source commercial products also use the Serverless model, including Confluent Cloud, MongoDB Atlas, Snowflake, Databricks, and so on.

At the 2022 Cloud Habitat Conference, Zhang Jianfeng, President of Alibaba Cloud Intelligence, announced that the core cloud products are fully serverless

Alibaba Cloud is the first cloud manufacturer in China to provide Serverless computing services. Function computing is a FaaS product, which is an event driven fully hosted computing service. Users only need to write and upload code, and function computing will automatically prepare computing resources, run code in an elastic and reliable way, and provide complete observability, greatly simplifying the development, operation and maintenance process.

👉 Function calculation FC: https://www.aliyun.com/product/fc?

The Serverless application engine SAE is the first application oriented Serverless PaaS platform in the industry. It shields the complexity of the underlying IaaS and Kubernetes, provides a zero code transformation, more cost-effective, and more efficient application hosting solution, and helps users achieve serverless single Web applications, micro service applications, and scheduled tasks.

Alibaba Cloud leads the industry in introducing the Serverless container service ASK. Based on the Elastic Container Instance (ECI), Alibaba Cloud can expand the capacity of 2000 pods in one minute, reducing the threshold for Kubernetes, allowing users to focus more on applications rather than managing the underlying infrastructure.

In addition, in 2020, Alibaba Cloud's open-source Serverless Devs will become the first cloud native lifecycle management platform supporting mainstream Serverless services/frameworks in the industry, and will officially enter CNCF Sandbox in September 2022, becoming the first selected Serverless tool project in the industry.

Serverless driven R&D mode upgrade

Each Serverless service is the output of the manufacturer's capabilities in the field, and promises reliability, flexibility, performance and other capability indicators. Therefore, they are high-quality application building blocks. For example, Alibaba Cloud Object Storage (OSS) service carries EB level massive data, promises 11 nines of data reliability, 99.95% of availability, and diversified hierarchical data storage and processing capabilities. Alibaba Cloud message queue RocketMQ, tempered by the peak of 11 trillion messages, promises 10 nines of data reliability and 99.95% of availability.

These cloud services have obvious advantages in flexibility, reliability, etc. compared with the system built by customers based on open source software. As manufacturers launch more and more Serverless services in storage, computing, middleware, big data and other fields, and these services are tightly integrated through event driven and other ways, the cloud has gradually become a super platform for application construction and operation. The application R&D mode has also been upgraded to assembly type R&D.

The figure below shows the comparison between traditional R&D mode and assembly type R&D:

Serverless is the best paradigm for cloud computing to benefit thousands of industries

From emerging Internet startups to traditional enterprises building large-scale software, Serverless architecture and assembly based R&D can be used.

Based on the Serverless architecture, a nine person team of start-ups created wolai, a new form of document application for team collaboration. The whole application uses cloud services such as function computing, object storage, table storage, and Redis. The system consists of multiple loosely coupled functions. The release or bug fix of each function point only needs to update the relevant functions. The function release rhythm is shortened to several hours, and it can be realized from model selection to online in a short time. The operation and maintenance cost will be saved by 700000-800000 yuan annually, and the cost will be reduced by 50%.

Pumpkin Movie has increased its new users by 1 million within one hour due to a hot movie. The traffic surge has caused the interruption of website services for a while, and the temporary cloud expansion cannot meet the huge traffic in a timely manner. We learned from our mistakes and began to transform to Serverless. It took seven days to complete the upgrade and make a qualitative leap. Finally, the development efficiency is increased by 70%, the cost is reduced by 40%, and the operation and maintenance efficiency is increased by 10 times, so as to calmly deal with sudden traffic.

Century Lianhua was the first new retail representative to test Serverless. It found that the old architecture of Dajieshi could not meet the problem of traffic explosion. It put the membership system, trading system, payment system, etc. on Alibaba Cloud function computing. It said goodbye to the large-scale business supported by expanding machines. The promotion preparation time was shortened from weekly to hourly, and the R&D and operation and maintenance efficiency were improved by 30%, while the cost was reduced by 40%. It really turned the promotion activities into a normal state;

The cloud has catalysed the change of R&D mode. Serverless architecture and assembly type R&D enable customers to do less and gain more, which will certainly become the mainstream R&D paradigm on the cloud.