Network protocols often suffer from undetected performance degradations due to inadequate testing and a lack of efficient and accurate evaluation tools across diverse network configurations. Current methodologies face challenges like pre-modeling, oversimplifications, and focusing on limited failure modes, making them costly, impractical, or imprecise.We propose Mahak, a novel black-box framework that reconstructs the empirical performance of a given protocol throughout the multidimensional configuration space using an active learning-guided sampling strategy, without prior modeling or knowledge of the internal algorithm of the protocol. Applied to state-of-the-art Internet Congestion Control (e.g., BBR2, Sage, Orca) and Adaptive Bitrate Streaming protocols (e.g. Pensieve, BOLA, RobustMPC), Mahak explores less than 0.1% of the configuration space, and achieves up to a 12.5× reduction in mapping error compared to interpolation-based methods.By systematically identifying the complete empirical performance surface, rather than focusing on a single prominent failure, Mahak uncovers issues that would otherwise remain hidden. This end-to-end mapping equips protocol designers, QA engineers, and network operators with actionable data-driven insights across diverse metrics and configurations, supporting more reliable deployments.

In this paper, our primary objective is to showcase that the application of machine learning techniques extends beyond network protocol design. We aim to demonstrate that performance assessment of network protocols, a vital aspect of improving network infrastructures and developing better protocol designs, can be modernized through the utilization of machine learning. As a step towards this goal, we have designed and introduced Mahak, the first tool that harnesses active learning techniques to automate the performance assessment of congestion control schemes. Mahak actively learns to optimize the evaluation process of congestion control schemes so that they can generate their performance maps over a desired space without exhaustively testing them in every scenario. Mahak treats schemes under the test as black boxes. This protocol-agnostic aspect of Mahak enables users to directly assess the performance of the actual implementation of a protocol instead of their over-simplified mathematical models or simplified simulated versions.

Traffic processing on cloud-scale bandwidths has given rise to a new type of network structure, comprising a large number of highly-structured virtual entities working in close harmony. This structure, which we call a virtual sub-network, might be in need of migration, for reasons of load-balancing, maintenance, and disaster prevention. In this paper, we argue that the common migration schemes are not adequate for the complexity of this task. Therefore, we present Qanat, a migration system specifically optimized for the live migration of a virtual sub-network in its entirety to a different physical location. We show how Qanat employs widely-used techniques, such as traffic prioritization, buffering, and network tunnels, to overcome the main issues of live migration. In the paper, we categorize the main challenges of the migration task, provide an analytical study of Qanat’s algorithms, and measure its performance metrics through large-scale simulations. We conclude that Qanat can efficiently and transparently migrate virtual sub-networks and can provide a useful tool for system administrators.

In this work, for the first time, we demonstrate that computers can automatically learn from observing the heuristic efforts of the last four decades, stand on the shoulders of the existing Internet congestion control (CC) schemes, and discover a better-performing one. To that end, we address many different practical challenges, from how to generalize representation of various existing CC schemes to serious challenges regarding learning from a vast pool of policies in the complex CC domain and introduce Sage. Sage is the first purely data-driven Internet CC design that learns a better scheme by harnessing the existing solutions. We compare Sage's performance with the state-of-the-art CC schemes through extensive evaluations on the Internet and in controlled environments. The results suggests that Sage has learned a better-performing policy. While there are still many unanswered questions, we hope our data-driven framework can pave the way for a more sustainable design strategy.

Due to the dynamic nature of traffic, networks must rapidly adapt to changing conditions. This is especially true in the context of the control plane which must ensure continuous and seamless operation. Switch migration, the process of changing the controller associated with a switch, is an important tool in facilitating this goal. In this work, we study the problem of minimizing the overall time to migrate a set of switches. We examine the problem subject to constraints on controller resources and QoS groups. We show that the problem is NP-hard and provide heuristic algorithms for solving large instances in practice. Through extensive experiments, we demonstrate that the heuristics achieve performance close to optimal while reducing the running time by several orders of magnitude.

Can we instead of designing yet another new TCP algorithm, design a TCP plug-in that can enable machines to automatically boost the performance of the existing/future TCP designs in cellular networks? We answer this question by introducing DeepCC. DeepCC leverages advanced deep reinforcement learning (DRL) techniques to let machines automatically learn how to steer throughput-oriented TCP algorithms toward achieving applications' desired delays in a highly dynamic network such as the cellular network. We used DeepCC plug-in to boost the performance of various old and new TCP schemes including TCP Cubic, Google's BBR, TCP Westwood, and TCP Illinois in cellular networks. Through both extensive trace-based evaluations and real-world experiments, we show that not only DeepCC can significantly improve the performance of TCP schemes, but also after accompanied by DeepCC, these schemes can outperform state-of-the-art TCP protocols including new clean-slate machine learning-based designs and the ones designed solely for cellular networks.

These days, taking the revolutionary approach of using clean-slate learning-based designs to completely replace the classic congestion control schemes for the Internet is gaining popularity. However, we argue that current clean-slate learning-based techniques bring practical issues and concerns such as overhead, convergence issues, and low performance over unseen network conditions to the table. To address these issues, we take a pragmatic and evolutionary approach combining classic congestion control strategies and advanced modern deep reinforcement learning (DRL) techniques and introduce a novel hybrid congestion control for the Internet named Orca1. Through extensive experiments done over global testbeds on the Internet and various locally emulated network conditions, we demonstrate that Orca is adaptive and achieves consistent high performance in different network conditions, while it can significantly alleviate the issues and problems of its clean-slate learning-based counterparts.

Since, current widely available network protocols/ systems are mainly throughput-oriented designs, meeting stringent delay requirements of new applications such as virtual reality and vehicle-to-vehicle communications on cellular network requires new network protocol/system designs. C2TCP is an effort toward that new design direction. C2TCP is inspired by in-network active queue management designs such as RED and CoDel and motivated by lack of a flexible end-to-end approach which can adapt itself to different applications' QoS requirements without modifying any network devices. It copes with unique challenges in cellular networks for achieving ultra-low latency (including highly variable channels, deep per-user buffers, self-inflicted queuing delays, and radio uplink/downlink scheduling delays) and intends to satisfy stringent delay requirements of different applications while maximizing the throughput. C2TCP works on top of classic throughput-oriented TCP and accommodates various target delays without requiring any channel prediction, network state profiling, or complicated rate adjustment mechanisms. We have evaluated C2TCP in both real-world environment and extensive trace-based emulations and compared its performance with different TCP variants and state-of-the-art schemes including PCC-Vivace, Google's BBR, Verus, Sprout, TCP Westwood, and Cubic. Results show that C2TCP outperforms all these schemes and achieves lower average delay, jitter, and 95th percentile delay for packets.