In this article, we investigate the real-world capability of the multi-link operation (MLO) framework -- one of the key MAC-layer features included in the IEEE 802.11be amendment -- by using a large dataset containing 5 GHz spectrum occupancy measurements on multiple channels. Our results show that when both available links are often busy, as is the case in ultra-dense and crowded scenarios, MLO attains the highest throughput gains over single-link operation (SLO) since it is able to leverage multiple intermittent transmission opportunities. As for latency, if the two links exhibit statistically the same level of occupancy, MLO can outperform SLO by one order of magnitude. In contrast, in asymmetrically occupied links, MLO can sometimes be detrimental and even increase latency. We study this somewhat unexpected phenomenon, and find its origins to be packets suboptimally mapped to either link before carrying out the backoff, with the latter likely to be interrupted on the busier link. We cross validate our study with real-time traffic generated by a cloud gaming application and quantify MLO's benefits for latency-sensitive applications.

翻译：本文利用包含多信道5 GHz频谱占用度测量的大规模数据集，对IEEE 802.11be修正案中的关键MAC层特性之一——多链路操作（MLO）框架的实际能力进行了研究。结果表明，在超密集和拥挤场景下两条可用链路频繁繁忙时，MLO能够利用多个间歇性传输机会，相比单链路操作（SLO）获得最高的吞吐量增益。在延迟方面，若两条链路的占用度在统计上呈现相同水平，MLO可将延迟降低一个数量级。相反，在非对称占用的链路中，MLO有时会产生负面影响甚至增加延迟。我们研究了这一略显反常的现象，发现其根源在于执行退避前数据包被次优地映射至某条链路，而退避过程极可能在更繁忙的链路上被中断。我们通过云游戏应用生成的实时流量对研究进行交叉验证，并量化了MLO对延迟敏感型应用的性能优势。