{"id":111865,"date":"2023-11-30T15:10:24","date_gmt":"2023-11-30T13:10:26","guid":{"rendered":"https:\/\/prohoster.info\/blog\/novosti-interneta\/dostupna-platforma-optimizaczii-trafika-libreqos-1-4"},"modified":"2023-11-30T15:10:24","modified_gmt":"2023-11-30T13:10:26","slug":"dostupna-platforma-optimizaczii-trafika-libreqos-1-4","status":"publish","type":"post","link":"https:\/\/prohoster.info\/en\/blog\/news\/dostupna-platforma-optimizaczii-trafika-libreqos-1-4","title":{"rendered":"The Traffic Optimization Platform LibreQoS 1.4 is Available","gt_translate_keys":[{"key":"rendered","format":"text"}]},"content":{"rendered":"

The release of the LibreQoS platform 1.4 has been published, designed for the fair distribution of available bandwidth among users and to reduce the negative effects caused by packet buffering (Bufferbloat) by network equipment. The platform can be used by providers or administrators of private networks to optimize traffic flows, maintain delays at a minimal level, and allocate bandwidth according to priorities. The project's code is written in C, Python, and Rust, and is distributed under the GPLv2 license. The project is being developed under the leadership of Dave Taht, co-founder of the Bufferbloat project, creator of the CeroWrt distribution, and author of numerous RFCs related to queue management. <\/p>\n

LibreQoS reduces latency and increases the reliability of interactive sessions, online gaming, online learning platforms, VoIP traffic, and video calls under heavy network load, such as when users are downloading movies in multiple streams or engaging in torrent activity (LibreQoS addresses the issue of video call stuttering when someone on the same network starts downloading 4K video). The implementation of LibreQoS reduces the peak bandwidth available to a single user but significantly decreases delays and fairly distributes resources among all data exchange participants. In a conducted test, the use of LibreQoS reduced data reception latency from 106 ms to 9 ms, and transmission latency from 517 ms to 23 ms, at the cost of lowering continuous download speed from 74 Mbps to 25 Mbps and upload speed from 29 Mbps to 8 Mbps. <\/p>\n

At the core of LibreQoS is the use of the CAKE (Common Applications Kept Enhanced) queue management system and the fq_codel (Fair Queuing Controlled Delay) packet scheduler, along with the use of eBPF and XDP (Express Data Path) for executing handlers at the network driver level with direct access to DMA buffers. The CAKE algorithm is designed to replace and simplify the complex hierarchy of queue management disciplines, capable of extracting maximum possible throughput and providing minimal latency even on the slowest service provider links and while operating on low-power devices. <\/p>\n

LibreQoS also provides tools for tracking delays between sending a request and receiving a response (RTT, round-trip time), linked to individual users, access points, and websites. A web interface has been developed for analyzing the status, allowing visual assessment of network traffic, monitoring changes in load and delays, and identifying the most active users. Flexible hierarchical traffic limiting schemes can be created and integrated with UISP and Splynx for mapping topologies and clients. <\/p>\n

LibreQoS is installed on server<\/a>, positioned between the provider's border router and the local network's main router. One server with LibreQoS can handle traffic shaping for thousands of users, for example, server<\/a> a 16-core Xeon Gold CPU is sufficient to manage ISP client traffic with a capacity of 11 gbit\/s. <\/p>\n

In the new version: <\/p>\n