/Patent Tracker
MicrosoftGranted patentChips & hardwareInterest 3/10

Microsoft patents faster data loading by skipping decompression steps

Microsoft has been granted a patent for a method that reads compressed data directly into memory without first decompressing it. The system converts the data's compression format on the fly using a technique called radix clustering to organize the data efficiently. This process aims to reduce the time it takes for a computer to start processing queries from stored files.

Why it mattersThis technical improvement could make database systems and cloud services respond to user requests slightly faster by optimizing how data is moved from hard drives to active memory.

In the filing’s own words

Example solutions read compressed data from a file in persistent storage, directly into an in-memory store without decompression. Some examples transcode the compression scheme used by the file into the compression scheme used by the in-memory store whereas, in other examples, the in-memory store is compatible with the compression scheme used by the file. In some examples that use transcoding, radix clustering is used to speed up compression dictionary transcoding. The radix clustering minimizes cache misses, thereby increasing the efficiency of memory access. These approaches significantly improve cold start times when responding to queries.

What exactly they claim (claim 1)

1. A radix clustering system comprising: a processor; and a computer-readable medium storing instructions that are operative upon execution by the processor to: read compressed data from a file, the compressed data having a first storage compression scheme with a first storage compression dictionary and a second storage compression dictionary; hash entries in the first storage compression dictionary and the second storage compression dictionary; sort the hashed entries by their associated hash values; update a hash table with the hashed entries; and based on at least the updated hash table, transcode the first and second storage compression dictionaries into an in-memory compression dictionary.

Facts

Company
Microsoft (Microsoft Technology Licensing, LLC)
Official title
Reading compressed data directly into an in-memory store
Patent number
US12743447B2
Application number
19023104
Filed
Jan 15, 2025
Granted
Sep 22, 2026
Inventors
4
Claims
20

Classified under

G06F 16/285 Physics & computingG06F 16/1744 Physics & computingG06F 16/2255 Physics & computing

More from Microsoft in Chips & hardware

Summaries are simplified from the filing’s own abstract and first claim. A patent filing is not a product announcement.
Patent data and drawings: U.S. Patent and Trademark Office (USPTO). Nebula is not affiliated with or endorsed by the USPTO.