The system developed by Huawei to manage its cloud involves dynamic traffic allocation optimized by machine learning. It was a response to the increased demand for services during the pandemic, especially in live streaming and online meetings.
Huawei's approach included an array of methodologies like continuous optimization, integer programming, graph theory, scheduling, and network flow problem solving, along with state-of-the-art machine learning algorithms. The focus was on minimizing network bandwidth costs associated with live streaming services.
The authors identified the surge in live streaming users, driven by the COVID-19 pandemic, as a key factor in the need for the dynamic traffic allocation system. Companies shifted digital assets to the cloud, leading to sudden high demand and quality expectations from cloud service providers.
Huawei's solution involved dissecting and optimizing the usage of bandwidth resources at different stages of live streaming services. By breaking down the process into distinct modules, the system aimed to minimize costs related to network bandwidth while meeting the escalated demand.
[
add
]
[
|
|
...
]