We use Alibaba Cloud services to enable serverless workflows. For example, we use Function Compute to handle services at the API layer and run common and custom applications on Function Compute. We also use Tablestore to store data and Serverless Workflow to complete the underlying implementation of all automated workflows. We use downstream data lakes for data analysis and use Log Service to store logs. This reduces costs for serverless workflows, improves service elasticity, and contributes to the smooth and concurrent executions of automated workflows during peak traffic hours.

About Qinglin Cloud

Qinglin Cloud, a professional Internet service company, provides technical and product-related consulting and services to help enterprises facilitate digital transformation and business expansion. Qinglin Cloud has provided mature and complete solutions that cover multiple industries. These solutions have served hundreds of millions of users.

Challenge

To adapt to the rise of enterprise automation, such as robotic process automation (RPA) and IFTTT, a customer decided to automate its workflows. To achieve this goal, the customer entrusted Qinglin Cloud to build a platform similar to IFTTT. In this case, personnel from other companies, including technical and non-technical personnel, can edit the business logic by dragging and dropping modules on web pages or mini programs. After enterprises customize or purchase the applications, they can edit the business logic based on workflows and flexibly change policies. This automates marketing, user operations and growth, enterprise workflows, and other workflow-related businesses.

Pain points of the customer include:

  • The structure of the applications that run automated workflows is complex. This increases the difficulty in application development and requires high costs and multi-type resources.
  • The service stability of the applications that run automated workflows is difficult to guarantee. The downtime of a single node may interrupt the entire service.
  • Workflow status management is complex. The customer finds it difficult to troubleshoot issues, such as execution failures in long-running workflows and memory leaks.
  • The issues caused by distributed architectures of applications that run automated workflows are challenging, and task scheduling requires external components.
  • Multi-tenant resource isolation may fail, and the concurrent workflow executions may conflict with each other.
  • The customer has to bear significant maintenance fees for server resources and queue messages.

Why Alibaba Cloud

Qinglin Cloud uses Alibaba Cloud services to enable serverless workflows. For example, it uses Function Compute to handle services at the API layer and runs common and custom applications on Function Compute. It also uses Tablestore to store data and Serverless Workflow to complete the underlying implementation of all automated workflows. Qinglin Cloud uses downstream data lakes for data analysis and uses Log Service to store logs. This reduces the cost for serverless workflows, improves service elasticity, and contributes to the smooth and concurrent executions of automated workflows during peak traffic hours.

Architecture

Serverless Workflow simplifies the development and maintenance of applications that run automated workflows, and reduces the development cycle and costs. The sophisticated Serverless Workflow model of Alibaba Cloud provides complete logic support for workflow automation at the initial roll-out stage of applications. A large number of automated workflows run in parallel. Managed services from Alibaba Cloud help the customer ensure the smooth and stable running of services during peak traffic hours. This also improves customer experience. The application model of Serverless Workflow ensures resource isolation in multi-tenant mode, and avoids conflicts in concurrent executions of automated workflows. Serverless Workflow and Function Compute significantly reduce costs, increase development efficiency by dozens of times, and also contribute to the expansion of business.

Looking Forward

In the current version, custom applications that run automated workflows require external services. The later versions will still allow you to deploy custom applications to Function Compute, and are expected to enable workflow automation for more business scenarios.

Looking Forward

  • Function Compute: A secure and stable, elastically scaled, O&M-free, pay-as-you-go, serverless computing platform.
  • Serverless Workflow: Visualization, O&M-free orchestration, and Coordination of Stateful Application Scenarios.
  • Tablestore: A fully managed NoSQL cloud database service that enables storage of a massive amount of structured and semi-structured data.
  • Log Service: An all-in-one service for log-type data.