"The development team explains: Advanced Filtering addresses a common problem in vector search - combining vector search with filtering (...) DiskANN Advanced Filtering solves this by combining filter and search into one operation - while the graph of vectors is traversed during the vector search, each vector is also checked for filter predicate match, ensuring that only the correct vectors are retrieved."
"HorizonDB pairs upstream compatibility with disaggregated storage designed to make Postgres really shine. We're pushing as much replication and durability work as we can into the storage layer leaving more CPU for PostgreSQL to run queries and transactions. It reduces commit latencies, make failovers fast and predictable, and scale reads across replicas while integrating with Azure security and AI tooling."
Azure HorizonDB is a managed, Postgres-compatible cloud database service that supports up to 3,072 vCores across primary and replica nodes and auto-scaling shared storage up to 128 TB with claimed sub-millisecond multi-zone commit latencies. The service integrates Microsoft Foundry models via an AI Model Management feature and claims up to 3x faster vector search using DiskANN advanced filtering versus HNSW. DiskANN Advanced Filtering combines filter and vector search into one traversal to return only vectors matching predicate filters. Disaggregated storage offloads replication and durability to the storage layer, freeing CPU for PostgreSQL, reducing commit latencies, speeding failovers, and scaling reads while integrating with Azure security and AI tooling. New database management improvements are being introduced.
Read at InfoQ
Unable to calculate read time
Collection
[
|
...
]