1. The computer that ran the revolution

Robert A. Heinlein’s The Moon is a Harsh Mistress features a computer that has grown so complex that it develops something like personality and starts making decisions that surprise its operators. The desktop’s network will not develop personality, but it may develop behavior that surprises ground controllers if it is allowed to route around failures on its own.

This entry works through what changes if the desktop’s network can heal itself.

2. Operations from the ground become operations from the cells

Today, a network change is a planned event. Ground controllers define the new topology, verify it against constraints, and upload the configuration. With self-reconfiguration, the cells themselves decide when to change routes. Ground is informed, consulted on policy, and called in for exceptions, but the milliseconds of a failover are handled locally.

This shifts the operations model. Controllers stop being router administrators and become policy administrators.

3. Failure recovery becomes faster

In a fixed network, a failed link or cell requires ground to detect the problem, diagnose it, design a new route, and upload the fix. That takes minutes at best and hours at worst. In a self-reconfiguring network, the cells can detect a dead neighbor and reroute in seconds or less. The desktop becomes more resilient to the small failures that are normal in a long mission.

4. New failure modes appear

Autonomy introduces its own risks:

  • A cell could be isolated by a transient fault and declare itself failed before ground can assess it.
  • A routing loop could form and saturate the network.
  • A new cell with bad software could be accepted into the federation before its credentials are verified.
  • The network could optimize for local health in ways that violate global constraints, such as power budgets or security policies.

These risks are manageable, but they require careful protocol design.

5. What this changes

  • Self-reconfiguration shifts the operations model from ground-in-the-loop to cells-in-the-loop.
  • It speeds up failure recovery but introduces new classes of autonomy risk.
  • The protocols must distinguish between local decisions and global policy.
  • The next entry will ask what the desktop must provide to make this possible.