Flent

Abstract

Running network performance experiments on real systems is essential for a complete understanding of protocols and systems connected to the internet. However, the process of running experiments can be tedious and error-prone. In particular, ensuring reproducibility across di erent systems is di cult, and comparing di erent test runs from an experiment can be non-trivial.In this paper, we present a tool, called Flent, designed to make experimental evaluations of networks more reliable and easier to perform. Flent works by composing well-known benchmarking tools to, e.g., run tests consisting of several bulk data ows combined with simultaneous latency measurements. Tests are speci ed in source code, and several common tests are included with the tool. In addition, Flent contains features to automate test runs, collect relevant metadata and interactively plot and explore datasets.We showcase Flent's capabilities by performing a set of experiments evaluating the new BBR congestion control algorithm, using Flent's capabilities to reproduce experiments both in a controlled testbed and across the public internet. Our evaluation reveals several interesting features of BBR's performance.

TC-RAN: A Programmable Traffic Control Service Model for 5G/6G SD-RAN

“…There, packets are handled in a FIFO manner, and thus, one greedy flow can monopolize the queue, negatively affecting the remaining flows tagged with the same QFI, since no contention or fairness mechanism is foreseen. In fact, from the MAC scheduler's perspective 2 DiffServ and QFI use 6 bits i.e., 2 6 = 64 classes there exists a group of different queues with diverse QFIs, where the ordering of arrival within the queues needs to be respected, forming a partially ordered set or poset [15]. Additionally, the MAC scheduler lacks any ability to segregate the packets within an RLC queue, as their contents may have been ciphered for security reasons at the Packet Data Convergence Protocol (PDCP).…”

“…As shown in Fig. 1, the throughput and Round Trip Time (RTT) of two contemporary softwarized 5G RAN stacks in SA mode under the Real-Time Response Under Load (RRUL) test provided by flent [2], present delays in the order of hundreds of milliseconds, which exceeds the latency envisioned by 3GPP for low-latency scenarios [3]. During the first 5 seconds in the RRUL test, only the UDP and ping flows are active, and thus, we can observe that the no-load delay of the system is around 10 ms, a suitable value for time sensitive over the top (OTT) applications [4].…”

Transmission Control Protocol and Active Queue Management together against congestion: cross-comparison through simulations

“…The paper shows that, in a general access network, the performance of ARED is similar (and in particular cases even better) to that of its competitors. With regard to TCP, this related work investigated only the standard TCP New Reno as the TCP deployed in the end nodes; this is also the case with another study on the stability of TCP/AQM systems proposed by Sun et al 37,38 In a different work, 39 based on Hoiland-Jorgensen et al, 40 the focus is again on AQM and, in general, on the queueing discipline available at the bottleneck; the authors describe the mechanism involving the technology of the access network and the bottleneck characteristics considering only TCP Cubic as the TCP congestion control. A similar discussion is reported by Anelli et al, 41 in which a novel AQM is proposed and compared to some of the last AQMs proposed in the literature but, unfortunately, tests are carried considering only TCP New Reno as the congestion control.…”

Uma Abordagem de Fatiamento de Rede entre Múltiplos Sistemas Autônomos

“…Segundo, foi experimentado a sobrecarga de tempo em termos de atraso fim-afim causado pelo processamento dos pacotes em uma fatia multidomínio por meio da ferramenta Flent (HØILAND-JØRGENSEN et al, 2017). Adicionalmente, foi verificado o desempenho do fatiamento de rede em múltiplos domínios providos pelo NASOR com o cenário não orientado ao fatiamento de rede -notadamente um baseline -e com a abordagem baseada em VPN.…”