Skip to main content

​
What is Intel QAT Engine?

Intel QuickAssist Technology (QAT) OpenSSL Engine is a plugin that provides cryptographic acceleration for OpenSSL-based applications. It leverages Intel QuickAssist Technology to offload compute-intensive cryptographic operations from the CPU, delivering significant performance improvements for secure communications.

Hardware Acceleration

Offload crypto operations to dedicated QAT hardware accelerators on Intel platforms

Software Optimization

Utilize optimized CPU-based implementations using AVX-512 and Intel instruction sets

Seamless Integration

Works as a drop-in plugin for OpenSSL through the engine or provider interface

Broad Algorithm Support

Accelerate RSA, ECDH, ECDSA, AES-GCM, ChaCha20-Poly1305, and more

​
Architecture Overview

The QAT Engine sits between your application and OpenSSL, intercepting cryptographic operations and routing them to the most appropriate acceleration path:

​
Dual Acceleration Paths

The QAT Engine supports two distinct acceleration paths that can be used independently or together for maximum flexibility.

​
QAT_HW: Hardware Acceleration

Uses dedicated Intel QuickAssist Technology hardware accelerators present in:
  • Intel Xeon Scalable Processors with C62X chipset
  • Intel QAT 4xxx Series adapters
  • Intel Atom C3000 and P5900 processors
Supported algorithms include:
  • Asymmetric: RSA, DH, DSA, ECDH, ECDSA, X25519/X448, SM2
  • Symmetric: AES-CBC-HMAC-SHA, AES-GCM, AES-CCM, ChaCha20-Poly1305, SM4-CBC
  • Key Derivation: TLS PRF, HKDF
  • Hashing: SHA3, SM3

​
QAT_SW: Software Acceleration

CPU-based acceleration using optimized instruction sets on 3rd Gen Intel Xeon Scalable Processors and newer:
  • Leverages AVX-512 and vector instructions
  • Uses Intel Crypto Multi-buffer library
  • Intel IPSec Multi-buffer library for AES-GCM
Supported algorithms include:
  • Asymmetric: RSA (2048/3072/4096), ECDH (X25519, P-256, P-384, SM2), ECDSA (P-256, P-384, SM2)
  • Symmetric: AES-GCM (128/192/256), SM4-CBC/GCM/CCM
  • Hashing: SM3

​
Key Capabilities

Non-blocking asynchronous job handling allows applications to submit multiple crypto operations concurrently, maximizing throughput in multi-threaded environments. The engine supports OpenSSL’s async job infrastructure with configurable async_jobs parameters.
OpenSSL pipelining enables parallel processing of large data buffers by splitting them into smaller chunks. The QAT Engine supports up to 32 pipelines with 16KB per pipeline for chained cipher operations.
The heartbeat feature monitors QAT device health and automatically falls back to software implementations if hardware becomes unavailable, ensuring continuous operation even during device resets or failures.
Native provider interface support for OpenSSL 3.0 alongside the traditional engine interface, providing a modernized integration path for new applications.
Built with FIPS 140-3 Level-1 certification requirements including self-tests, integrity checks, and CMVP/CAVP compliance for regulated environments.
Works with OpenSSL, Tongsuo (BabaSSL), and BoringSSL, providing flexibility for different deployment scenarios and application requirements.

​
Use Cases

The QAT Engine is ideal for workloads that require high-throughput cryptographic operations:

Web Servers

NGINX and Apache handling thousands of TLS connections per second

Load Balancers

HAProxy terminating SSL/TLS at scale

VPN Gateways

IPsec and SSL VPN concentrators

Content Delivery

CDN edge servers with HTTPS traffic

​
Performance Benefits

Hardware acceleration with QAT can deliver:
  • 10-100x throughput improvement for RSA operations
  • 2-10x improvement for ECDH/ECDSA operations
  • 2-5x improvement for AES-GCM bulk encryption
  • Significantly reduced CPU utilization, freeing cycles for application logic
Actual performance gains depend on your specific workload, hardware configuration, algorithm choice, and async_jobs tuning. See the performance testing section for benchmarking guidance.

​
Next Steps

Quickstart

Get up and running with QAT Engine in minutes

Installation Guide

Detailed installation instructions for your platform

Core Concepts

Learn about QAT_HW and QAT_SW acceleration

API Reference

Explore the complete API documentation

Build docs developers (and LLMs) love

Get started for freeTalk to us