Abstract:We consider the problem of designing a packet-level congestion control and scheduling policy for datacenter networks. Current datacenter networks primarily inherit the principles that went into the design of Internet, where congestion control and scheduling are distributed. While distributed architecture provides robustness, it suffers in terms of performance. Unlike Internet, data center is fundamentally a "controlled" environment. This raises the possibility of designing a centralized architecture to achieve better performance. Recent solutions such as Fastpass and Flowtune have provided the proof of this concept. This raises the question: what is theoretically optimal performance achievable in a data center? We propose a centralized policy that guarantees a per-flow end-to-end flow delay bound of $O$(#hops $\times$ flow-size $/$ gap-to-capacity). Effectively such an end-to-end delay will be experienced by flows even if we removed congestion control and scheduling constraints as the resulting queueing networks can be viewed as the classical reversible multi-class queuing network, which has a product-form stationary distribution. In the language of Harrison et al., we establish that baseline performance for this model class is achievable. Indeed, as the key contribution of this work, we propose a method to emulate such a reversible queuing network while satisfying congestion control and scheduling constraints. Precisely, our policy is an emulation of Store-and-Forward (SFA) congestion control in conjunction with Last-Come-First-Serve Preemptive-Resume (LCFS-PR) scheduling policy.

A Concise Queuing Model for Controller Performance in Software-Defined Networks.

Performance Evaluation of OpenFlow-based Software-Defined Networks Based on Queueing Model

An Active Queue Management Scheme Using Neural Network Based Predictive Control

Performance Evaluation for Sdn Deployment: an Approach Based on Stochastic Network Calculus

Modeling and Optimization of Packet Forwarding Performance in Software-Defined WAN

Performance Modelling and Optimization of Controller Cluster Deployments in Software-Defined WAN

Achieving Real-Time Quality of Service with Software Defined Networking

Predictive Control applied Queuing Strategy in Networked Control Systems

A Multipath Resource Updating Approach for Distributed Controllers in Software-Defined Network

A Performance Analytical Approach Based on Queuing Model for Network-on-Chip

Simulation based performance analysis of networked control systems with resource constraints

Cutting Long-Tail Latency of Routing Response in Software Defined Networks.

Modeling and optimization of the data plane in the SDN-based DCN by queuing theory

On the Queue Dynamics of Multiuser Multichannel Cognitive Radio Networks

Traffic Modeling and Performance Evaluation of SDN-based NB-IoT Access Network.

Admission control in cognitive radio networks with finite queue and user impatience

A Dormant Multi-Controller Model for Software Defined Networking

How Powerful Switches Should be Deployed: A Precise Estimation Based on Queuing Theory

Centralized Congestion Control and Scheduling in a Datacenter

On Performance Modeling for MANETs under General Limited Buffer Constraint

Performance Evaluation and Optimization of Hierarchical Routing in SDN Control Plane