FlowMesh: A Service Fabric for Composable LLM Workflows

TL;DR

FlowMesh is a scalable, cost-effective service fabric that optimizes composable LLM workflows by decomposing tasks, sharing work across users, and dynamically managing resources, significantly reducing costs and energy use.

Contribution

It introduces FlowMesh, a novel multi-tenant service fabric that efficiently executes and optimizes complex LLM workflows with fine-grained operators and a global control plane.

Findings

Up to 3.8x cost reduction compared to baseline solutions

2.0x lower energy consumption

Maintains efficiency under dynamic and failure-prone conditions

Abstract

AI deployment increasingly resembles a pipeline of data transformation, fine-tuning, and agent interactions rather than a monolithic LLM job; recent examples include RLHF/RLAIF training and agentic workflows. To cope with this shift, we propose FlowMesh, a multi-tenant service fabric that executes and optimizes these workloads as one shared service instead of isolated pipelines. It decomposes workflows into fine-grained operators with recorded lineage, enabling de-duplication of work across users and batching requests on the same hardware while preserving per-workflow provenance. A global control plane maintains a cluster-wide pool of ready operators and uses a single utility function to pick both the batch and the worker, balancing throughput, cost, and data locality on heterogeneous GPUs. The data plane is an elastic fleet of stateless workers backed by a content-addressable store, enabling rapid, automatic scale-out, safe retry after preemption, and portability across managed clusters such as Kubernetes and geo-distributed GPU marketplaces such as Vast.ai. Compared with baseline solutions, FlowMesh achieves up to 3.8x cost reduction and 2.0x lower energy usage, provides a similar or better latency profile, and remains efficient under dynamic and failure-prone conditions.