Can't find new article about YouTube architecture, the last one is from 2012 - " 7 Years Of YouTube Scalability Lessons In 30 Minutes ", and it says:
(找不到有关YouTube体系结构的新文章，上一篇是2012年的文章-“ 30分钟内YouTube的可伸缩性教训7年 ”，其中说：)
YouTube is not asynchronous, everything is blocking.
(YouTube不是异步的，一切都在阻塞。)
Not-very-popular videos are stored at file system.
(不太受欢迎的视频存储在文件系统中。)
I understand that YouTube have I/O concurrency on RAID level (faster I/O), but still some videos may take up to 10 seconds. (我了解YouTube在RAID级别具有I / O并发（更快的I / O），但是某些视频可能仍需要10秒钟的时间。)
There are so many reactive frameworks now and it's hard to believe YouTube is still blocking. (现在有许多反应式框架，很难相信YouTube仍在阻止。)
Why YouTube use blocking requests?
(为什么YouTube使用阻止请求？)
Is it still blocking at 2019? (它在2019年是否仍然受阻？)
ask by VB_ translate from so 与恶龙缠斗过久,自身亦成为恶龙；凝视深渊过久,深渊将回以凝视…
