ElasticRec: A Microservice-based Model Serving Architecture Enabling Elastic Resource Scaling for Recommendation Models

TL;DR

ElasticRec is a microservice-based architecture for recommendation system model serving that enables elastic resource scaling and high memory efficiency, significantly reducing deployment costs.

Contribution

It introduces a novel elastic, microservice-based architecture with utility-based resource allocation for RecSys, improving resource utilization and reducing costs.

Findings

3.3x reduction in memory allocation size

8.1x increase in memory utility

1.6x reduction in deployment cost

Abstract

With the increasing popularity of recommendation systems (RecSys), the demand for compute resources in datacenters has surged. However, the model-wise resource allocation employed in current RecSys model serving architectures falls short in effectively utilizing resources, leading to sub-optimal total cost of ownership. We propose ElasticRec, a model serving architecture for RecSys providing resource elasticity and high memory efficiency. ElasticRec is based on a microservice-based software architecture for fine-grained resource allocation, tailored to the heterogeneous resource demands of RecSys. Additionally, ElasticRec achieves high memory efficiency via our utility-based resource allocation. Overall, ElasticRec achieves an average 3.3x reduction in memory allocation size and 8.1x increase in memory utility, resulting in an average 1.6x reduction in deployment cost compared to state-of-the-art RecSys inference serving system.