![]() My other worry is that PoPs can deploy DoS detection heuristic that are tuned for standard web pages and a bandwidth intensive page may trigger the rules and have the traffic throttled. PoPs can't certainly cache everything forever and they can be more eager to remove large files and in such case CDN will be reducing performance by adding an additional HTTP level hop to the origin server (and likely a few IP level hops). I worry how long large files are going to be cached by CDN PoPs. ![]() The primary use case for CDN is latency reduction for serving small files (CSS, JS, images) and such traffic has different characteristic and requires different optimizations than high-throughput traffic. I did not find any general purpose CDN that would explicitly encourage the use of their product for large files nor a CDN that would disclose any information how large payloads are handled by their infrastructure. I'm pondering using standard CDNs for serving bandwidth intensive payloads, but I have some doubts whether this is a right technical decision. Good luck to CloudFlare and congrats to Bitmovin. Using correct terminology and customising stock Bitdash player would have helped at the start of such ambitious endeavour. Bandwidth-heavy customers do want to take advantage of rapidly commoditising technology and falling prices and are pushing for multi-CDN strategy. They just distribute whatever and however many bytes origin server has for them (give or take some convenience features around it).ĬDNs are scrambling to provide compelling features to increase stickiness, usually with limited success when it comes to video. However, while Akamai surely would love to bill for more bytes per minute of video, no one is asking them. It’s amusing to hear an argument that CDNs are hampering adoption of better video compression from a CDN company. Multiple resolutions, adaptive bitrates, dynamic packaging for different versions of HLS/DASH/Smooth are pretty much a must-have for any video solution these days, free or paid. What’s the deal with ‘lossless compression’? I feel like they use the term very differently from how the industry uses it. Surely, this piece of the pie is not as big as many had hoped for, but there’s definitely more than a 1000 companies that think their video content is worth a better platform than YouTube or Facebook, however they define “better”. With YouTube and Facebook owning an overwhelming majority of video hosting market, what’s left is shared between Brightcove, The Platform, Ooyala, Bitmovin, Vimeo, Vidyard, Wistia and a whole ton of smaller players and in-house solutions (ffmpeg -> _any cdn_ -> video.js). CloudFlare is entering a very crowded space.
0 Comments
Leave a Reply. |