1. Don’t panic, but keep a checksum

Douglas Adams’ The Hitchhiker’s Guide to the Galaxy returns as the recalled work because its advice is as useful for storage as it is for travelers: know what you have, keep a backup, and do not rely on a single copy of anything important. A self-healing federation needs its own towel — a few simple, reliable primitives that the cells can use to detect damage, locate redundancy, and repair fragments.

This entry lists what the desktop would have to provide for federated storage healing to work.

2. Physical provisions

Healing is only possible if the data is spread across independent hardware:

  • Cell-level redundancy: each fragment must be stored on a different cell so that one cell failure does not wipe out multiple copies.
  • Independent power domains: a power fault on one cell should not corrupt fragments on another.
  • Network connectivity: surviving cells must be able to read fragments from each other and write reconstructed fragments to a replacement or spare cell.
  • Spare capacity: the federation must reserve enough space and bandwidth to absorb a failed cell and rebuild its data.

3. Identity, addressing, and naming

The cells must know what they are storing and where the other copies live:

  • Content-addressed or globally unique object identifiers: each fragment is named by what it contains, not by which cell holds it.
  • Location registry: the federation maintains an eventually consistent map of which cells hold which fragments.
  • Version vectors or timestamps: so cells can decide which copy is current after a partition.
  • Persistent identity: cells and storage devices keep their identity across reboots and swaps.

4. Integrity and repair

The cells need mechanisms to notice and fix damage:

  • Checksums or hashes per fragment: to detect bit rot, transfer errors, and corruption.
  • Periodic scrubbing: cells read their own fragments, verify checksums, and report mismatches.
  • Erasure coding or replication: enough redundancy to reconstruct lost fragments from the remainder.
  • Repair throttling: healing runs at a rate that does not overwhelm the network, power, or thermal budget.
  • Audit logs: every detected fault and every repair is logged for ground review.

5. Policy and security

A federation that moves data around must still obey rules:

  • Redundancy policy: ground defines how many fragments, on how many cells, in which physical locations.
  • Access control: not every cell can read every fragment.
  • Quarantine: a misbehaving or failing cell can be excluded from healing until it is diagnosed.
  • Verification: repairs are verified before the old fragments are released.

6. What this changes

  • Federated healing requires investment in redundancy, naming, integrity, and policy infrastructure.
  • Many of these provisions are worth making even for ground-managed backups.
  • The next entry will decide whether the desktop should design for this capability.