Correct remote layout and privacy docs for hashed snapshot keys (closes #67)
check / check (pull_request) Failing after 1s

The remote layout and threat model in three documents described plaintext
snapshot IDs as directory names and misattributed the observable backup
time to those IDs. In fact `RemoteSnapshotKey` names each metadata
directory (and the manifest `snapshot_id`) with a one-way double SHA-256
hash of the human ID, so hostname and snapshot name are not observable;
the backup time is, via the plaintext manifest timestamp, an accepted
design property (issue 81).

Document the derivation once in `docs/REPOSTRUCTURE.md` with a worked
example; README, ARCHITECTURE and DATAMODEL now show the hashed layout and
link to it. Rewrite the privacy section to state what the unencrypted
manifest really exposes. Fix two code comments that claimed the public
bytes hide the timestamp. Docs and comments only; no behaviour change.

Model: opus-4-8
This commit is contained in:
2026-09-21 13:00:54 +00:00
parent d2a0510cb4
commit 978fef8b2c
6 changed files with 69 additions and 28 deletions
+14 -4
View File
@@ -344,7 +344,7 @@ both are set.
├── blobs/
│ └── <aa>/<bb>/<full_blob_hash>
└── metadata/
└── <snapshot_id>/
└── <remote-key>/
├── db.zst.age # Encrypted binary SQLite database
└── manifest.json.zst # Unencrypted blob list (for pruning)
```
@@ -355,8 +355,18 @@ both are set.
* `manifest.json.zst` is an unencrypted compressed JSON blob list, enabling
pruning without the private key
Snapshot IDs follow the format `<hostname>_<snapshot-name>_<RFC3339-timestamp>`
(e.g. `server1_home_2025-06-01T12:00:00Z`).
Snapshot IDs follow the human-readable format
`<hostname>_<snapshot-name>_<RFC3339-timestamp>` (e.g.
`server1_home_2025-06-01T12:00:00Z`), but this ID is never written to the
destination store in plaintext. Each snapshot's metadata directory is named
with its `<remote-key>`, a one-way double SHA-256 hash of the ID, so a listing
of the store reveals no hostname or snapshot name. The backup time is not
hidden: manifest.json.zst carries a plaintext timestamp, and object
modification times are visible at the storage layer regardless. For example,
`server1_home_2025-06-01T12:00:00Z` is stored under
`metadata/17f97bcde958748af076b926af59823943db59e80ce7170b40f124dfa28f64aa/`.
See [docs/REPOSTRUCTURE.md](docs/REPOSTRUCTURE.md#remote-key-derivation) for the
derivation.
### data flow
@@ -373,7 +383,7 @@ Snapshot IDs follow the format `<hostname>_<snapshot-name>_<RFC3339-timestamp>`
**restore:**
1. Download and decrypt `metadata/<snapshot_id>/db.zst.age`
1. Download and decrypt `metadata/<remote-key>/db.zst.age`
2. Open the binary SQLite database
3. Query files (optionally filtered by paths)
4. Download and decrypt required blobs