Amazon patents database system that splits data into hot and cold tiers
Amazon has filed a patent for a database architecture designed to store data linked with multi-dimensional vectors. The system divides storage into three levels: a 'hot' tier for fast writing, a 'cold' tier for fast reading, and a middle tier to balance both. When a user searches, the system checks all three levels and combines the results to find the closest matches.
In the filing’s own words
A vector-labeled database, which stores data in association with multi-dimensional vectors, comprises a hot tier which is optimized for writes, a cold tier which is optimized for reads, and a medium tier which is optimized to balance reads and writes. Incoming transactions are executed at the hot tier, and moved to the medium and cold tier in batches. Reads for the vector-labeled database search all tiers, and merge the results of the searches of the individual tiers.
What exactly they claim (claim 1)
1. A system, comprising: one or more first computing devices configured to implement a cold tier for a vector-labeled database, wherein the vector-labeled database stores data items and associated vector labels for the data items; and one or more second computing devices configured to implement a hot tier for the vector-labeled database; wherein: the cold tier is read-optimized, relative to the hot tier, to retrieve a given one or more of the data items based on a given search vector; the hot tier is write-optimized, relative to the cold tier, to perform transactions to store incoming data items and associated vector labels, the one or more first computing devices and the one or more second computing devices are further configured to, in response to receiving a query comprising a search vector, identify a first set of vectors from the cold tier based on a search using the search vector and identify a second set of vectors from the hot tier based on a search using the search vector, and from a merged set of the identified first and second sets of vectors determine a subset of vectors based on respective distances to the search vector to be returned in response to the query; and the cold tier is updated in batches, wherein a batch includes data items associated with vectors originally written to the hot tier.
Facts
- Company
- Amazon (Amazon Technologies, Inc.)
- Official title
- MULTI-TIERED VECTOR DATABASE ARCHITECTURE
- Publication number
- US20260288687A1
- Application number
- 19085873
- Filed
- Mar 20, 2025
- Published
- Sep 24, 2026
- Inventors
- 3
- Claims
- 20
Classified under
G06F 16/2237 Physics & computingG06F 16/24542 Physics & computingMore from Amazon in AI & software
- Amazon patents system that cites sources for AI answers Sep 29, 2026
- Amazon patents tech to swap words in movies using actor's voice Sep 8, 2026
- Amazon patents ML tool to add graphics to live sports from moving cameras Sep 15, 2026
- Amazon patents shopping assistant that remembers your search filters Oct 6, 2026
- Amazon patents system where AI agents split tasks and talk to each other Oct 20, 2026
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.