TL;DR
An experimental evaluation of ALTQ/CBQ demonstrating its sensitivity to a wide range of parameters and link layer driver design issues is attempted and suggestions for effective deployment in real networks are made.
Multipath QUIC aggregates the bandwidth available to multihomed clients. Path and stream scheduling policies are known to affect performance but rely on static, general-purpose approaches that fail to meet application requirements. We present a framework to request novel proxy-based services that require scheduling in network topologies with multiple paths to an HTTP/3 proxy. It repurposes the Extensible Prioritization Scheme (EPS) to convey tailored schedulers at the connection level. With our own open-source MASQUE implementation, Nada, we demonstrate that interplay by selecting a custom stream-aware scheduler at the proxy. A reproducible measurement campaign shows that such client-proxy cooperation with EPS as a platform enables Quality of Service to be requested for specific flows. Our contribution lays the groundwork for adaptive solutions that dynamically optimize for desired behaviors, such as handling path heterogeneity, without involving the target server.
The proposed Multi-Path Multi-Level Feedback Queueing (MP-MLFQ) leverages the spatial diversity and regularity of DCNs to realize a scheduler with numerous logical priority levels while occupying as low as 2 physical priority queues within network switches.
Transmission Control Protocol (TCP) congestion control plays a vital role in maintaining the quality of Internet communications. However, flows with different Round-Trip Times (RTTs) and Congestion Control Algorithms (CCAs) significantly degrade throughput fairness when sharing a bottleneck link. Especially in environments where CUBIC and BBR coexist, existing controls at intermediate nodes struggle to suppress bandwidth dominance by specific flows. This study proposes a method using the P4 programmable data plane to classify flows into independent queues based on two axes: RTT length and CCA type. Furthermore, applying Fair Queuing with Variable Holding Time (FQ-VHT) to each classified queue ensures equitable bandwidth distribution for each flow, even in environments where diverse communication characteristics coexist.
Network slicing is a key enabling technology for fifth-generation (5G) and beyond mobile networks, which enables operators to run multiple logical networks on top of common physical infrastructure while meeting heterogeneous quality-ofservice (QoS) requirements. In this paper, we summarize the design, implementation, and evaluation of a UDP-based virtual network slicing simulator modeling four slices in accordance with 3GPP service types: enhanced Mobile Broadband (eMBB), Ultra-Reliable Low-Latency Communications (URLLC), Massive Machine-Type Communications (mMTC), and a dedicated Best Effort slice. The simulator uses real UDP sockets on localhost in a multi-threaded framework and implements separate admission control per slice, supporting three configurable resourceborrowing modes—none, full, and controlled—in which the primary slices (eMBB, URLLC, mMTC) may borrow unused capacity from the Best Effort slice when their own allocation runs out. QoS metrics including throughput, packet loss ratio, one-way delay, and jitter are collected in a thread-safe manner per slice. Experimental results under controlled borrowing (60 s, 16 nodes, 3 MB per slice, 40% borrow cap) demonstrate activation of the borrowing mechanism with nonzero borrowed-in and borrowedout values, and illustrate the trade-off between primary-slice throughput and Best Effort protection.
Evaluation on representative workloads demonstrates that P2CS achieves performance comparable to in-network mechanisms while significantly reducing complexity and cost, and requires minimal software changes making it readily deployable in today's datacenter infrastructure.