High Throughput and Low Latency Compression Engine For SSD

Overview

Key Features

• Implements a LZ77 variant algorithm
• High performance capabilities, meaning the highest compression ratio for a given search buffer size. It delivers 1.5 times more compression compared to the most popular hardware compression algorithm, LZW.
• High throughput of 16Gbps is achieved in ASIC. However, with custom implementations the compression throughput can be as high as 64Gbps.
• Low latency is achieved by designing the engine in two pipe stages. The first pipe stage is referred to as the "Search" pipe in which the search for the longest match string takes place. The second pipe stage is the "Post Processor" used to output unmatched characters and length/offset pairs.

Benefits

• Advantages of compression in SSD and enterprise storage is well known. In terms of SSD, it allows for better Input/Output Operations Per Second (IOPS), more space, an increase in lifespan and more bandwidth. For enterprise storage, compression increases disk capacity and performance. The major challenge in designing compression / decompression for Solid State Devices (SSDs) and enterprise storage is the reduction in performance due to latency and throughput. This low latency and high throughput design helps resolve the issues listed above.

Applications

• SSD Controller
• Enterprise Storage

Deliverables

• RTL
• Testbench & Tests
• Documentation

Technical Specifications