What is Redis SDS and Why is it Better Than C Strings?

What is SDS?

SDS (Simple Dynamic String) is Redis's custom string implementation, designed specifically for high-performance scenarios. As Redis's most fundamental data structure, it underpins all key-value storage and list elements. Its design philosophy can be summarized in three keywords: safety, efficiency, and flexibility.

SDS Internal Structure

Each SDS contains three key metadata elements:

1. Length Field (len): Real-time tracking of current string length (8 bytes)
2. Free Space (free): Pre-allocated unused buffer size (8 bytes)
3. Character Array (buf): Flexible array for actual content storage

This design creates a three-part memory layout: header metadata + data area + reserved space. Here's an example storing "Hello":

len=5          ← Current string length
free=3         ← Available free space
buf=['H','e','l','l','o','\0','',''] ← Actual storage (includes null terminator and reserved space)

Core Advantages

1. Fast Length Access

While traditional C strings require traversing the entire array to get the length (O(n) time complexity), SDS directly reads the len attribute (O(1)). When accessing LIST and other data structures, this optimization can improve performance by over 100x.

2. Buffer Overflow Prevention

SDS's space pre-allocation strategy completely eliminates C string overflow risks:

• For strings < 1MB: Doubles capacity during expansion
• For strings ≥ 1MB: Adds 1MB extra space per expansion

This mechanism reduces memory reallocation frequency by over 50% when handling APPEND operations.

3. Binary Safety

SDS can store any binary data (including content with '\0' characters), enabling Redis to safely store:

• Binary image data
• Serialized Protobuf messages
• PDF file segments

In contrast, C strings truncate at '\0', causing data corruption.

4. C String Compatibility

While SDS functionally surpasses C strings, it maintains '\0' termination, allowing it to:

• Directly use C string library functions
• Seamlessly integrate with existing system APIs
• Reduce developer learning curve

5. Memory Optimization

Through lazy space reclamation, SDS doesn't immediately free memory when shortening strings. Instead, it tracks excess space in the free field. This space can be reused during future expansions, reducing memory allocations by 30-50%.

6. Type Classification

Redis implements 5 SDS types (sdshdr5~sdshdr64) for different string lengths, minimizing memory overhead through differentiated metadata headers:

SDS Applications in Redis

1. Key-Value Storage: All Redis keys are SDS type
2. List Elements: LIST structure elements rely on SDS
3. Persistence Buffer: AOF/RDB file writing uses SDS as buffer
4. Network Protocol: RESP protocol parsing is implemented with SDS

Related Resources:

• Practice Redis Interview Questions Like LeetCode
• sds.h
• sds.c