CMDRL: A Markovian Distributed Rate Limiting Algorithm in Cloud Networks

As cloud networks continue to evolve, network traffic has experienced an exponential increase. The network architecture is progressively adopting a distributed structure to address this challenge. This architecture extensively utilizes technologies like gateway clusters and Equal-Cost Multi-Path (ECMP) routing, enabling traffic from individual tenants to be routed through multiple pathways. As a result, distributed rate limiting (DRL) has emerged as an essential aspect. Nonetheless, the shift from centralized to DRL has encountered obstacles, with the associated algorithms grappling with simplicity, precision, and applicability issues. Consequently, our research seeks to reconceptualize the issue of DRL from a theoretical standpoint to discover a more holistic and efficacious solution.

In this paper, we introduce CMDRL, a novel Markovian DRL algorithm that conceptualizes the problem as a random walk on a graph and frames it within a Markov model premised on two widely accepted assumptions. We establish that the converging model exhibits a bounded mixing time, serving as a more comprehensive metric for assessing convergence velocity. The algorithm’s benefits include simplified input requisites, greater versatility across diverse contexts, and improved accuracy. Our evaluation results show that the algorithm attains a favorable convergence, evidenced by an L1-norm error of 0.063 after 50 iterations, demonstrates a rapid adaptive response with a 72.9% decrease in error within 20 iterations, and achieves a swift mixing time, requiring only 33 iterations to reach an error threshold of 0.05. Additionally, We evaluate the effectiveness of various optimization strategies for Markov chains in addressing the DRL problem.