Bring Your Own Objective: Inter-operability of Network Objectives in Datacenters

2026-02-10

AI Summary

Datacenter networks struggle to simultaneously satisfy the diverse multi-objective requirements of modern applications, including flow completion time, fairness, and low latency. To address this challenge, this work proposes DMart, a decentralized scheduling framework that models bandwidth as a competitive market, enabling applications to express their priorities through autonomous bidding and thereby natively coexist under multiple objectives. DMart employs a distributed, per-link, per-RTT auction mechanism that supports large-scale deployment with sub-microsecond latency overhead, eliminating the need for centralized schedulers, integer linear programming, or complex priority queues. Experimental results demonstrate that DMart matches pFabric’s short-flow completion times while reducing deadline miss rates by 2× and improving coflow completion times by 1.6× compared to Sincronia.

Technology Category

• Planning, Routing, and Scheduling: Scheduling under Uncertainty
• Search and Optimization: Distributed Search
• Constraint Satisfaction and Optimization: Distributed CSP/Optimization

Application Category

• Economics, Online Markets and Human Computation: Incentives in network design for Web infrastructures and ecosystems
• Systems and Infrastructure for Web, Mobile and WoT: Data management and stream processing for Web, mobile and wireless applications
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs

Abstract

Datacenter networks are currently locked in a"tyranny of the single objective". While modern workloads demand diverse performance goals, ranging from coflow completion times, per-flow fairness, short-flow latencies, existing fabrics are typically hardcoded for a single metric. This rigid coupling ensures peak performance when application and network objectives align, but results in abysmal performance when they diverge. We propose DMart, a decentralized scheduling framework that treats network bandwidth as a competitive marketplace. In DMart, applications independently encode the urgency and importance of their network traffic into autonomous bids, allowing diverse objectives to co-exist natively on the same fabric. To meet the extreme scale and sub-microsecond requirements of modern datacenters, DMart implements distributed, per-link, per-RTT auctions, without relying on ILPs, centralized schedulers, or complex priority queues. We evaluate DMart using packet-level simulations and compare it against network schedulers designed for individual metrics, e.g., pFabric and Sincronia. DMart matches the performance of specialized schedulers on their own"home turf"while simultaneously optimizing secondary metrics. Compared to pFabric and Sincronia, DMart reduces deadline misses by 2x and coflow completion times by 1.6x respectively, while matching pFabric short-flow completion times.

Problem

• datacenter networks
• network objectives
• performance diversity
• single-objective tyranny
• coflow

Innovation

• decentralized scheduling
• network interoperability
• bandwidth marketplace
• per-link auctions
• multi-objective optimization