Introduction

When matching a massive set of descriptors using classical brute-force methods, it becomes impossible to maintain low latency with a high request rate. Therefore, it is necessary to employ approximation techniques that trade off some accuracy for significant speed improvements. These techniques expedite the matching process by optimizing the data and creating an efficient index.

Luna Index Module (LIM) is designed to build efficient indices on a set of descriptors and perform matching on it. LIM consists of following components:

• Manager - server that processes index building tasks and coordinates indexers

• Indexer - builds indices

• Matcher - performs approximate nearest neighbor matching using indices built

LIM has the following applicability limitations:

1. The index becomes outdated as soon as new objects are actively created or deleted in the Luna Platform.

2. Index building is a time-consuming process and should not be done frequently.

3. Index building consumes a significant amount of resources and time.

4. The matcher continues to search against the outdated version of the index until it is rebuilt.