https://k4i.top/tags/distributed-systems/Recent content in Distributed-Systems on k4i's blogHugo -- gohugo.ioensky_io@outlook.com (K4i)sky_io@outlook.com (K4i)All content is subject to the license of <a rel="license noopener" href="https://creativecommons.org/licenses/by-nc-sa/4.0/" target="_blank">CC BY-NC-SA 4.0</a> .Wed, 17 Jun 2026 10:00:00 +0800https://k4i.top/posts/streaming-application-layer-design/Wed, 17 Jun 2026 10:00:00 +0800sky_io@outlook.com (K4i)Wed, 17 Jun 2026 10:00:00 +0800https://k4i.top/posts/streaming-application-layer-design/<p>When people first design a streaming interface, the natural question is:</p> <p><strong>If TCP already provides a reliable ordered byte stream, and HTTP/2 or HTTP/3 already supports long-lived multiplexed streams, why do we still need application-layer streaming design?</strong></p> <p>Because the transport layer only promises to move bytes. Real systems care about higher-level facts: which file do these bytes belong to? Is this chunk number 17? Can this operation be retried? What if the same chunk arrives twice? Can a download resume after disconnecting halfway through? Who slows down when the consumer cannot keep up?</p>K4ifeatured imagenetworkingstreamingdistributed-systemsbackendnotes