NegotiaToR: Towards A Simple Yet Effective On-demand Reconfigurable Datacenter Network

TL;DR

NegotiaToR addresses dynamic traffic in reconfigurable DCNs by introducing a minimalist, on-demand optical network design that uses in-band distributed scheduling. It employs a two-phase epoch with a simple, non-iterative NegotiaToR Matching algorithm, binary per-ToR requests, and data piggybacking to bypass scheduling delays for latency-sensitive mice flows, achieving substantially improved mice FCT and goodput on flat AWGR-based topologies. The approach emphasizes practicality and scalability, validating performance across parallel and thin-clos topologies, and showing robustness to parameter variation, fault conditions, and diverse workloads. Overall, NegotiaToR provides a low-complexity, high-performance path toward practical on-demand reconfigurable DCNs with direct one-hop data transmission.

Abstract

Recent advances in fast optical switching technology show promise in meeting the high goodput and low latency requirements of datacenter networks (DCN). We present NegotiaToR, a simple network architecture for optical reconfigurable DCNs that utilizes on-demand scheduling to handle dynamic traffic. In NegotiaToR, racks exchange scheduling messages through an in-band control plane and distributedly calculate non-conflicting paths from binary traffic demand information. Optimized for incasts, it also provides opportunities to bypass scheduling delays. NegotiaToR is compatible with prevalent flat topologies, and is tailored towards a minimalist design for on-demand reconfigurable DCNs, enhancing practicality. Through large-scale simulations, we show that NegotiaToR achieves both small mice flow completion time (FCT) and high goodput on two representative flat topologies, especially under heavy loads. Particularly, the FCT of mice flows is one to two orders of magnitude better than the state-of-the-art traffic-oblivious reconfigurable DCN design.

Figures (19)

• Figure 1: Flat topologies in AWGR-based optical DCNs.
• Figure 2: Each epoch in NegotiaToR comprises two phases, and reconfiguration only happens in the predefined phase.
• Figure 3: NegotiaToR Matching's workflow. (a) shows the workflow on the parallel network topology. (b) and (c) illustrate the variance in the GRANT step across two topologies, attributable to differing connection capabilities. On the parallel network topology, port ①-④'s grant priority is determined by a shared ring, whereas on thin-clos it's determined by port-specific rings. Once granted, the pointer is updated to prioritize the next source.
• Figure 4: Scheduling for different epochs is performed in a pipelined manner.
• Figure 5: Bypassing scheduling delay with unscheduled transmission in the predefined phase.