This project is under development, and not production ready.

Scaling and maintain file storage is one the most use case. Using AWS/GCS and etc not accessible by all users and some post process of files like converting will be difficult.

• Convert Video/Audio to standard HLS.
• tusd Upload support.
• Multi domains/buckets.
• CDN/Secure/Performance in mind design.
• POSIX Design. For independent any solution
• Shardable/Distributed Design.
• Thumbnail generation.

All environment variable must has prefix ASM_.

• Start with ASM_PUBLIC will expose as public configuration, otherwise consideration as secure/private variable.

This simple emoji show when to change the variables of adms;

• 🚀: Config this variable during launch time and NEVER change it.
• 🔧: Config of this could be change any time for changing application workload/behavior.
• 🚫: Config of this variable may cause service different behavior.
• 🔴: Config of this variable is for special use cases or testing, do not touch it if you're not knowing how that works.

We have server id 0, bucket localhost with file hash 2x7pLVtJRWkwaq that contain simple jpeg file;

• For best practices is good to use base62 of using mongodb ObjectID for easy integration. (eg 507f1f77bcf86cd799439011 is 2x7pLVtJRWkwaq)
• [FILE_DIR] use levels=1:2 hash of file like nginx proxy_cache_path level for gain more performance and backup in mind solutions for rsync.

Each deployment need bucket files that standard json file

[
  {
    // identifier of buckets
    "id": "localhost",
    // if active and new upload will be proceed otherwise just serve will work
    "active": true,
    // is private that each serve request need JWT token for access data, otherwise all files is public resource.
    "private": false,
    // list of domains that accepted by this bucket
    "domains": [
      "localhost",
      "127.0.0.1"
    ],
    // list of vod qualities that convert to HLS
    "vod": [
      "360p",
      "720p"
    ],
    // list of aod qualities that convert to HLS
    "aod": [
      "128k"
    ],
    // JWT HS256 secret for sign/verify upload token
    "upload_secret": "upload_secret",
    // JWT HS256 secret for delete or get info of server
    "management_secret": "management_secret",
    // JWT HS256 secret for sign/verify serve token
    "serve_secret": "serve_secret"
  }
]

• For thumbnail generation JWT token require for prevent DoS/DDoS of cpu usage, even server is public "private": false.
• Domains must unique across all buckets.

Consider you have sample example.com and you need store media files. We need scale media server to two servers.

Scaling and distribute requirement to calculate many parameters.

• How much you will need for post processors.
• How much need space for your media.

These server could be private or protected via any mechanism like vpn, wireguard, upstream secure or etc.

Consider we have public ip like 5.x.x.x.

up[N].example.com will accept the upload request and process the media and files. s[N].example.com will serve the content via cache in CDN/Reverse Proxy layer for increase your throughput.

Now you can upload and process in two separate nodes.

For increase high serving content you can add more s[N].example.com node and cache serving content.