Introduction to CDN System Architecture and Workflow

This article describes the basic components and workflow of a traditional Content Delivery Network (CDN), excluding complex logic such as grouping, layering, and P2P. It first outlines each CDN component and their relationships, then demonstrates a simple download example as an introductory guide.

System Composition

1.1 System Overview

CDN is a content distribution network system that includes storage and distribution of text information as well as binary data. Text information is stored and distributed via a real‑time database synchronization mechanism, while binary data uses a proprietary protocol.

The CDN service network abstracts the following components: a CDN content management service, a central resource management database, numerous service nodes distributed across various network regions, and several CDN business entry points.

1.2 CDN Logical Diagram

02 – CDN Components and Function Introduction

The CDN module diagram clearly shows several components, which are abstracted as the following services: entry service, core index service, core storage service, content management service, edge storage service, and content distribution service.

2.1 Entry Service

The entry service is the CDN system’s gateway; all requests first hit this service, which provides a redirection address to the client for subsequent requests.

2.1.1 Specific Functions

Store information related to the core index service.

Store information related to the content management service.

DNS resolution and redirection: based on client address and other data, the entry service directs the client to the appropriate service.

2.1.2 Routing Strategies

Client‑wide configuration routing strategy.

IP‑based region and ISP routing strategy.

Load‑balancing strategy.

Speed‑test strategy, etc.

2.2 Core Index Service

The core index service provides query capabilities for the CDN distribution network. When the distribution system needs to download or cache a new content file, it queries the index service to locate the storage server that holds the content.

2.2.1 Specific Functions

Provide queries for edge storage services when a file is not cached locally.

Store information about core storage services.

2.3 Core Storage

Core storage is the central storage module of the CDN, a distributed storage system with specific algorithms that holds all binary data, offering fault tolerance and load‑balancing capabilities.

2.3.1 Specific Functions

Connect to the core index service and register all content information; updates are sent to the index service on additions or deletions.

Download content: after obtaining the download address from the core index service, the core storage establishes a connection to retrieve the file.

Upload content: an interface allows saving content files into the CDN system.

Execute backup commands: if a requested file is absent locally, the core storage performs a backup operation.

2.3.2 Core Storage Backup Flowchart

2.4 Content Management Service

The content management service manages edge storage services. It collects load information from edge nodes and cache status of binary content to decide which edge storage should serve a request.

2.4.1 Specific Functions

Provide content queries for edge storage services.

Supply a list of available content distribution services when a download is requested.

Manage the group of edge storage nodes.

File cache strategy: determine placement of each content fragment on edge nodes and recompute load‑balancing when nodes fail or are added.

File redirection: return the edge storage location to lower‑level nodes requesting data.

2.5 Edge Storage Service

Edge storage acts as secondary storage, caching binary content downloaded from core storage or manually placed locally.

2.5.1 Specific Functions

Serve content to the content distribution service; if the requested file exists locally, provide it directly.

If the file is not present locally, request it from other servers, download it, then serve the content.

If the file does not exist in the current edge node, inform the distribution service of the correct storage location.

03 – CDN Workflow

3.1 Client Download Request Process

Client sends a download request to the entry service.

The entry service, based on client information and routing policies, returns a list of content management services.

Client contacts a content management service.

If the requested data resides within the management service’s group, it returns a list of content distribution services.

If not, the edge storage service queries the core index service for the data location.

The core index service identifies the core storage containing the data and returns it to the edge storage service.

Edge storage requests the data from the identified core storage; multiple edge nodes may download from a single core storage, with content stored as whole files in core storage and as shards in edge and distribution services.

The content management service finally provides the client with a list of content distribution services.

The client requests the actual data from a distribution service, which retrieves it from the appropriate edge storage node.

3.2 Client Download Process Flowchart