Scholarlink

Embedding-based recommender systems (RecSys) face critical bottlenecks in storing massive embedding tables (EMTs) and accelerating latency-sensitive lookup operations. While serverless computing and disaggregated memory architectures offer resource efficiency, existing solutions struggle with cold-start penalties, state transfer overheads, and rigid resource scaling for EMT-bound workloads. We present ServerlessRec, a serverless framework that integrates memory disaggregation with kernel-space RDMA-based remote memory mapping (rmap) and remote fork (rfork) to enable elastic EMT lookups. ServerlessRec decouples compute nodes (CNs) from memory nodes (MNs), dynamically autoscaling serverless functions on CNs for bursty lookups while storing EMTs on MNs. ServerlessRec improves latency-bounded throughput by 3.8 \(\times \) over DisaggRec, reduces resource waste by 62% vs. ElasticRec, which demonstrates how kernel-level memory disaggregation unlocks serverless advantages for EMT-bound applications.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

ServerlessRec: Fast Serverless Inference for Embedding-Based Recommender Systems with Disaggregated Memory

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Mingxuan Liu,
  • Jianhua Gu,
  • Tianhai Zhao

摘要

Embedding-based recommender systems (RecSys) face critical bottlenecks in storing massive embedding tables (EMTs) and accelerating latency-sensitive lookup operations. While serverless computing and disaggregated memory architectures offer resource efficiency, existing solutions struggle with cold-start penalties, state transfer overheads, and rigid resource scaling for EMT-bound workloads. We present ServerlessRec, a serverless framework that integrates memory disaggregation with kernel-space RDMA-based remote memory mapping (rmap) and remote fork (rfork) to enable elastic EMT lookups. ServerlessRec decouples compute nodes (CNs) from memory nodes (MNs), dynamically autoscaling serverless functions on CNs for bursty lookups while storing EMTs on MNs. ServerlessRec improves latency-bounded throughput by 3.8 \(\times \) over DisaggRec, reduces resource waste by 62% vs. ElasticRec, which demonstrates how kernel-level memory disaggregation unlocks serverless advantages for EMT-bound applications.