When I first started with Nostr back in 2023, IPFS sounded like a perfect match. But the more I played with it, the more convinced I became that IPFS is a broken and failed project. Nostr needs something simpler and more straightforward—and then hzrd149 (nprofile…wj75)'s Blossom came.
Ever since then, I've been dreaming about redundant decentralized storage with file fragmentation—something like what StorJ describes in their white paper:
"All data is encrypted and sharded using Reed Solomon erasure coding. The resulting pieces are distributed around the world. A typical large file is divided into 64 MB encrypted segments, and each encrypted segment is divided into 80 erasure-coded pieces—any 29 of these are sufficient to reconstruct the original file. Each of these 80 pieces is stored on a different drive globally, spread across diverse geographies, operators, power supplies, and networks."
In Blossom, this would be a killer feature. You would offer space on your server to others, and in exchange, get redundancy for your own files, or you would simply pay for it. File metadata could be published as a specific kind hosted by relays.
Aida
npub1aj…hdt20
2025-03-31 11:01:15
in reply to nevent1q…zhvz
Author Public Key
npub1ajdaw3j4g6aqv86alhn3df8jpulj0mxz3jjgwpm4uh598hc348gqthdt20Seen on
wss://nos.lolPublished at
2025-03-31 11:01:15Event JSON
{
"id": "efea6e601388d2daee65f960ed3f354b76ea36266c16d110fb68cf03bf86e914",
"pubkey": "ec9bd7465546ba061f5dfde716a4f20f3f27ecc28ca4870775e5e853df11a9d0",
"created_at": 1743418875,
"kind": 1,
"tags": [
[
"e",
"6515da2900d1cbff94d3b8989922a58a298873840466bdc4bb6cc39bed3647a1",
"",
"root"
],
[
"e",
"00001c5acadc3c7d99890eaf01cf123dfb4cdf8d25aa09da2dc1ba4c3f8d9a56"
],
[
"e",
"acf4e8bac971c0adb0de30f8f7c7593c9b18a2066a2ea730ea825b160a37e6f3",
"",
"reply"
],
[
"p",
"ee9d979e7e1418a9f7472baf82dc96082a0d4251b77bcb170a6ecec93fadff64"
],
[
"p",
"3bf0c63fcb93463407af97a5e5ee64fa883d107ef9e558472c4eb9aaaefa459d"
],
[
"p",
"ee9d979e7e1418a9f7472baf82dc96082a0d4251b77bcb170a6ecec93fadff64"
],
[
"p",
"45835c36f41d979bc8129830f2f5d92562f5343d6feddd6f30aa79480730f26e"
],
[
"p",
"266815e0c9210dfa324c6cba3573b14bee49da4209a9456f9484e5106cd408a5",
"",
"mention"
]
],
"content": "When I first started with Nostr back in 2023, IPFS sounded like a perfect match. But the more I played with it, the more convinced I became that IPFS is a broken and failed project. Nostr needs something simpler and more straightforward—and then nostr:nprofile1qqszv6q4uryjzr06xfxxew34wwc5hmjfmfpqn229d72gfegsdn2q3fgpr3mhxue69uhhxct5v4kxc6t5v5hxs7njvscngwfwvdhk6tcpzfmhxue69uhkummnw3e82efwvdhk6tcpz4mhxue69uhhyetvv9ujuerpd46hxtnfduhsk7wj75's Blossom came.\n\nEver since then, I've been dreaming about redundant decentralized storage with file fragmentation—something like what StorJ describes in their white paper: \n\"All data is encrypted and sharded using Reed Solomon erasure coding. The resulting pieces are distributed around the world. A typical large file is divided into 64 MB encrypted segments, and each encrypted segment is divided into 80 erasure-coded pieces—any 29 of these are sufficient to reconstruct the original file. Each of these 80 pieces is stored on a different drive globally, spread across diverse geographies, operators, power supplies, and networks.\"\n\nIn Blossom, this would be a killer feature. You would offer space on your server to others, and in exchange, get redundancy for your own files, or you would simply pay for it. File metadata could be published as a specific kind hosted by relays.\n",
"sig": "c0866760bcb6006cf97733f354bd75562bc490c5f7cc7d4cf667172bb7c04fac45087684c22381ed050580053dd3d9c925e1957fb148cc73f0f3a88900012e85"
}