Block-Oriented Lossless Decompressor for Data Center Network Throughput
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current data center storage requirements are increasing rapidly, and existing compression algorithms are limited in their ability to effectively compress all data types while supporting random access, which hinders the acceleration of network transfers and increases costs due to the need for frequent 'forklift upgrades' of networking equipment.
Innovation Solution
A block-oriented lossless compression system that encodes data using servers before writing it to storage and decodes it in the same rack where the application requests it, allowing for compressed data to be transferred across the network, reducing storage costs and increasing networking throughput without requiring a full upgrade of data center networking equipment.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If compression algorithms are used to reduce storage costs and increase transfer speed, then storage costs decrease and transfer speed increases, but compression algorithms have historically been restricted to certain limited use cases, have not performed well on all types of data, and have not supported random access into the compressed stream of data
Solution Approach 1:
The patent changes the fundamental parameter of data representation by implementing a custom compression scheme that transforms data into a format optimized for both compression ratios and random access capability. This allows the system to achieve compression benefits across all data types without the limitations of traditional algorithms.
Solution Approach 2:
The compression system segments data into manageable blocks that can be independently compressed and accessed. This segmentation enables random access into compressed data streams while maintaining compression efficiency, resolving the contradiction between compression benefits and access flexibility.
2Speed
If networking equipment throughput is increased to support faster data transfers, then data transfer speed increases, but such networking speed improvements require a 'forklift upgrade' of all IT infrastructure networking equipment components which is expensive
Solution Approach 1:
The patent substitutes mechanical/network hardware upgrades with a software-based compression solution. By implementing compression at the software level, the system achieves faster effective data transfer without requiring expensive hardware replacements, thereby resolving the contradiction between transfer speed and upgrade cost.
Solution Approach 2:
The system changes the data representation parameter through compression, reducing the amount of data that needs to be transferred across the network. This parameter change enables faster effective throughput without requiring proportional increases in networking equipment capacity.
3Productivity
If compression algorithms are used to reduce the amount of data transferred, then network transfer efficiency increases, but compression algorithms have not supported random access into the compressed stream of data
Solution Approach 1:
The compression system divides data into segments that maintain access flexibility. This segmentation strategy enables random access into compressed streams while preserving compression efficiency, simultaneously improving both transfer efficiency and operational ease.
Data Source
AI summary
A block-oriented lossless decompressor is used to decode encoded data fetched from storage that is subsequently transferred across a network in encoded (compressed) form. In examples described herein, applications executing at network nodes send GET requests, or similar messages, to storage systems, which can return compressed data this is decompressed in an intermediate node (between the storage node and the app), and can return compressed data that is decoded in the same network node in which the requesting application is running.