NAVIS: Concurrent Search and Update with Low Position-Seeking Overhead in On-SSD Graph-Based Vector Search

TL;DR

NAVIS is a novel on-SSD graph-based vector search system that significantly reduces position-seeking overhead, enabling higher concurrent search and update throughput with lower latency in large-scale high-dimensional workloads.

Contribution

NAVIS introduces layout-supported selective vector reads, dynamic entrance graph updates, and an entrance graph-aware cache to improve GVS efficiency and scalability.

Findings

Increases insertion throughput by up to 2.74x

Enhances search throughput by up to 1.37x

Reduces search latency by up to 25.26%

Abstract

On-disk graph-based vector search (GVS) has become the dominant approach for serving large-scale vector databases at high recall, but prior systems struggle to sustain concurrent search and update throughput on high-dimensional workloads. We find the main cause of this in position seeking, a full graph traversal that every update performs to locate neighbors before linking the new vector into the graph. Position seeking is fundamentally heavier than a search query, and its cost is further amplified by two systemic limitations of current GVS systems, packed layouts that couple every edge fetch to a full vector load, and a static entrance graph whose entry points drift away from newly inserted regions as updates accumulate. We present NAVIS, an on-SSD GVS system that drives down position-seeking overhead through (i) a layout-supported selective vector read that breaks the packed-page coupling without losing its locality benefits, (ii) a dynamic lightweight entrance graph update mechanism that reuses traversal information already produced by concurrent updates, and (iii) an entrance graph-aware edgelist cache that concentrates capacity on high-reuse paths near refreshed entry points. Across multiple large-scale high-dimensional benchmarks, NAVIS enhances average insertion throughput by up to 2.74x and average concurrent search throughput by up to 1.37x while reducing average search latency by up to 25.26%.