How Serverless Transforms Cloud Computing: Insights from CNBPS 2019

Introduction

During the second Cloud Native Technology Practice Summit (CNBPS 2019) in Beijing, Tencent Cloud architect Kong Lingfei delivered a talk titled “One Talk to Understand Serverless”. He introduced Serverless from a macro perspective, explained why it has become popular since 2018, and described how Tencent Cloud implements Serverless and the scenarios where it is applicable.

Serverless Overview

Serverless gained more attention than micro‑services and Kubernetes in 2016. The first product, AWS Lambda, appeared in 2014, followed by offerings from Microsoft, Google, and IBM in 2016. In China, Tencent Cloud launched SCF in 2017, Alibaba Cloud introduced its own Serverless product, and Tencent Cloud later released TCB (WeChat Mini‑Program Cloud Base) in 2018 and Serverless 2.0 (TSF Serverless) in 2019.

What Is Serverless?

Serverless does not mean the absence of servers; the underlying servers are managed by the cloud provider. It is an architectural paradigm rather than a framework or library. Developers focus solely on business logic, without worrying about CPU, memory, or databases.

Evolution of Cloud Computing Layers

The diagram shows four layers: On‑Premise, IaaS, PaaS, and the topmost FaaS layer where Serverless resides. As the stack moves upward, the responsibilities of developers shrink, allowing smaller teams to deliver features faster.

Serverless Technical Forms

Tencent Cloud currently offers three Serverless forms:

Cloud Function : Event‑driven functions that can be triggered by API Gateway, CKafka, CMQ, COS, etc.

TSF Serverless : An HTTP‑service style where the service process stays resident, supporting long‑connections, low latency, Spring Cloud, and Service Mesh.

Tencent Cloud Base (TCB) : A PaaS‑like solution for Mini‑Program development that bundles cloud functions, object storage, and databases via SDKs.

Cloud Function Component Architecture

The architecture consists of several layers:

Infrastructure layer: Docker and lightweight VMs (Serverless 1.0 uses Docker, Serverless 2.0 uses lightweight VMs) provide fast cold‑start (milliseconds).

Resource management layer: Cluster monitoring, automatic scaling, high‑availability (dual‑active clusters), and security protection.

Authentication & Authorization layer: Ensures function security.

Access layer: Handles trigger dispatch and function execution.

Orchestration layer: Manages workflow scheduling.

User layer: Developers only need to write business code and interact with databases, storage, or other services.

DevOps Tooling for Serverless

Beyond compute resources, Tencent Cloud provides a suite of tools to support development and operations:

IDE support: VS Code plugin, Web IDE, and CLI for deployment, debugging, and configuration.

API & SDK: Enables automation and custom integration.

DevOps workflow: Supports coding, building, packaging, deploying, testing, and releasing.

Product integration: Seamless connection with Git repositories, API Gateway, ServerlessDB, etc.

Logging: Console view with request IDs, and optional export to Tencent Cloud Log Service for persistent storage and advanced querying.

Application Scenarios

Cloud Functions can serve as back‑ends for browsers, mobile apps, and Mini‑Programs. Different triggers enable use cases such as WebSocket services, message processing, stream computation, and event notifications.

Customer Case: Tencent Maps

In this event‑function scenario, user interactions with Tencent Maps generate data that is streamed into Kafka, triggering a cloud function. The function processes the data and writes results back to Kafka, from where another process stores them in HBase and Elasticsearch for later querying.

TSF Serverless Use Case – BFF Layer

When an app spans Android, iOS, and Web, maintaining separate back‑ends for each client is costly. Introducing a Backend‑For‑Frontend (BFF) layer decouples front‑end and back‑end, handling authentication, logging, and data aggregation. TSF Serverless can host the BFF as an HTTP service, allowing front‑end engineers to implement full‑stack features without managing underlying infrastructure.

Conclusion

Serverless represents the ultimate direction of cloud computing by abstracting away infrastructure, enabling developers to focus on business value, reducing personnel requirements, and accelerating product iteration. Tencent Cloud’s multi‑form Serverless offerings, comprehensive tooling, and ecosystem integrations illustrate how the industry is moving toward a “no‑infrastructure” development paradigm.