Youtube has multiple datacenters located around the world. Each of these contains a lot of servers and storage.
This in turn is copied to multiple datacenters to protect from data loss.
Youtube also uses caching services that storage popular videos near the edge (at ISPs) the speed up performance and reduce bandwidth usage. Akamai is one such common caching service that Youtube uses, they also use Google CDN.
It depends on what level of abstraction you’re talking at.
At the physical level, videos are stored all over the place, split between and replicated across many and magnetic hard disk drives and solid state flash drives spread across the planet in data centers.
There are many layers of abstraction for storage above that though. No one at YouTube is addressing speicific inodes of specific disks on specific machines when it comes to video storage.
Rather YouTube uses an internal distributed blobstore similar to [Google Cloud Storage](https://cloud.google.com/storage) or Amazon S3. This blobstore is built on various other storage primitives / abstractions. One abstraction or layer is called [Colossus](https://cloud.google.com/blog/products/storage-data-transfer/a-peek-behind-colossus-googles-file-system), Google’s cluster-level filesystem.
Ultimately, YouTube stores its videos in a blobstore, but that blobstore is an abstraction which is really storing data using other lower level services which ultimately end up storing bytes on physical disks.
However, when *you* receive a YouTube video, you’re not getting the same video that’s stored there. You interact with YouTube through a content delivery network, which will replicate and cache content close to the edge, which has its own storage strategies for that.
When u upload video on YT, it is compressed and stored on servers. Server is filed with hard drives, where video is stored. 1 server can hold how knows how many TB of storage. For example lets say 1000tb per server, imagine whole buildings filled only with servers (10.000 servers times 1000tb) it comes to allot of memory. Imagine that on scale of who knows how many servers in the world and there is your answer hehe
Latest Answers