Coloc improves query speed by co-locating multimodal data in storage
Cross-Modal Data Co-Locality for Heterogeneous Storage Systems
Distributed, Parallel, and Cluster Computing
Summary
Many computer systems store different types of data separately, like images and text, even when they’re used together. This can slow down search queries because the system has to fetch data from different places. The authors created Coloc, a tool that groups related data closer together based on how often they’re accessed at the same time. This arrangement makes searches faster and reduces slow data movement, without changing how the data is searched or what the searches return.
What this means in practice
- •For cloud infrastructure teams: Optimize placement of multimodal vectors across storage tiers to reduce latency and data transfer costs in cloud deployments.
- •For multimedia application developers: Improve real-time search performance in applications using combined image and text embeddings by better co-locating data.
Authors
Abdullah Al-Mamun
Abstract
Distributed vector stores increasingly serve multimodal workloads whose representations are jointly accessed but independently partitioned and placed. As a result, cross-modal physical locality remains largely unoptimized. Multimodal queries can incur remote accesses and cross-tier data movement even when each modality's ANN search remains efficient. We present Coloc, a storage-management layer that makes cross-modal access locality a first-class data-layout signal. A Cross-Modal Access Graph combines semantic affinity, co-access, and temporal behavior. A movement-aware optimizer places frequently co-accessed vectors under capacity and topology constraints. A migration manager adapts placement as workloads shift by weighing locality gains against migration cost. A confidence-gated prefetcher hides remaining slow-tier latency. Coloc operates above existing ANN indexes without modifying embeddings, search, or retrieval semantics. We deploy Coloc on a multi-tier AWS testbed spanning GPU, NUMA-DRAM, NVMe, and S3. Coloc substantially reduces data movement and query latency while leaving ANN retrieval unchanged. Under a scheduled workload shift, its migration manager matches the locality of periodic global reorganization while moving 5.9x less data and completing 3.9x faster.