This results in unbalanced load across all CPU cores, which leads to slowness.īecause of this request-process pinning effect, requests that do CPU heavy or blocking IO tasks can slow down other requests. The NGINX worker (process) architecture has operational drawbacks for our use cases that hurt our performance and efficiency.įirst, in NGINX each request can only be served by a single worker. Architecture limitations hurt performance For some limitations, we optimized or worked around them.
Over the years, our usage of NGINX has run up against limitations. Let’s dig in on why we chose to replace our legacy service and how we developed Pingora, our new system designed specifically for Cloudflare’s customer use cases and scale. This proxy service powers our CDN, Workers fetch, Tunnel, Stream, R2 and many, many other features and products. Today, we’re focusing on a different part of the equation: the service that proxies traffic between our network and servers on the Internet. In the past, we’ve talked a lot about how browsers and other user agents connect to our network, and we’ve developed a lot of technology and implemented new protocols (see QUIC and optimizations for http2) to make this leg of the connection more efficient. Many Cloudflare customers and users use the Cloudflare global network as a proxy between HTTP clients (such as web browsers, apps, IoT devices and more) and servers. We could no longer get the performance we needed nor did NGINX have the features we needed for our very complex environment.
It was great for many years, but over time its limitations at our scale meant building something new made sense. Today we are excited to talk about Pingora, a new HTTP proxy we’ve built in-house using Rust that serves over 1 trillion requests a day, boosts our performance, and enables many new features for Cloudflare customers, all while requiring only a third of the CPU and memory resources of our previous proxy infrastructure.Īs Cloudflare has scaled we’ve outgrown NGINX. This is not only due to the delay and bottleneck between the monitoring point and the client but also due to the client mobile access network.Post Syndicated from Yuchen Wu original Introduction Our findings show that the accuracy of the QoE monitoring decreases with the distance of the PoP to the client. We investigate the influence of different points of presence (PoP) in the cloud and mobile network on the performance of the VNF for monitoring video buffer and QoE. To this end, we implement a VNF to analyze video flows in the network by using deep packet inspection. In this work, we investigate the feasibility of deploying a virtual network function (VNF) for video buffer and QoE monitoring on the Amazon Web Service cloud. With the paradigm of network function virtualization, network operators are able to deploy such a video monitoring function in the cloud. This allows reacting to quality degradation to improve the service. To address this problem, network operators need a mechanism to monitor the quality of experience (QoE) perceived by the user. However, a rapid increase in video traffic and users poses challenges for network operators maintaining user expectation. Because of the variety of video content on the Internet, a potential market is emerging for video providers. The fast growth of video streaming is responsible for a huge amount of traffic over the past few years.
0 kommentar(er)
