What would happen if a telegraph network formed a closed loop? A message would no longer merely travel from sender to receiver. If each station continuously observed and retransmitted it, the message could circulate indefinitely.
Modern computers store information by encoding data into distinguishable physical states. Hard disk drives use magnetic states, solid-state drives use trapped electrical charge, and optical media encode information through physical or optical variations in a material. These technologies have achieved extraordinary storage densities and read-write speeds. Their weakness, however, becomes apparent when we consider preservation across long timescales.
No conventional storage device preserves information indefinitely. Hard drives suffer mechanical failure and degradation. Solid-state drives experience charge leakage and wear. Optical media deteriorate, and even experimental archival materials remain vulnerable to physical destruction. Large-scale archives compensate for these limitations through redundancy, error detection, maintenance, and periodic migration to new storage devices.
This raises a question: why must long-term data preservation depend on maintaining a persistent physical storage medium at all? Instead of preserving information by keeping it stationary within matter, could we preserve it by keeping it continuously in motion?
Data in Motion is Weird
We define data as information represented through distinguishable states according to an interpretation scheme. A hard drive represents data using magnetic states that can be detected by its read head. An SSD represents data using electrical charge states. A CD, likewise, encodes data using microscopic physical structures that affect reflected light. The point stands that data is represented through distinguishable states.
Likewise, data can also be transmitted by radio, light, or other forms of electromagnetic radiation. In this case, information is encoded into measurable properties of a signal. Most applications of data transmission occur over terrestrial distances. Because electromagnetic radiation travels at the finite speed of light, these transmissions still experience propagation delay, but the delay is often negligible from a human perspective.
Across astronomical distances, however, propagation delay becomes significant. Communication with interplanetary spacecraft can take minutes or hours. Mars, depending on its position relative to Earth, is several light-minutes away. Voyager 1 is so distant that a radio signal transmitted from Earth takes almost a day to reach it. This finite propagation time creates an interesting property: while a signal travels between two distant points, the information encoded within it remains physically in transit.
Consider two regenerative communication relays. Relay A is located near Earth, while Relay B is located near Mars. Relay A begins with a copy of an archive and transmits the encoded data toward Relay B. Depending on the relative positions of Earth and Mars, the transmission may spend several minutes travelling through interplanetary space before reaching B. When Relay B receives the transmission, it does not preserve the archive in long-term storage. Instead, it receives the signal, reconstructs the encoded data, corrects transmission errors, and retransmits it toward Relay A.
The archive therefore begins a cycle:
The information is preserved not by remaining stationary within a storage medium, but by repeatedly propagating between regenerative relays. As a result, we cannot use this method of data storage for rapid access. One possible use of this method is data archival. Data can only be retrieved, read, or written when enough of the archive reaches a probe and becomes reconstructable.
The issue, however, is that using only two probes makes the architecture extremely fragile. Although error-correcting codes like Reed–Solomon can work effectively, using only two probes for data storage is unsafe due to the lack of redundancy. If a single probe becomes inactive, the propagation loop breaks. The transmitted data continues travelling outward at the speed of light. Any replacement probe launched afterward must travel slower than light and therefore cannot catch the escaping signal. Unless another receiver is already positioned to intercept the transmission, the data is lost forever into space.
Errors that remain undetected, or data that is incorrectly reconstructed, could also be retransmitted and propagate through subsequent cycles. So only two probes for data storage is considerably riskier.
Constellations & Relativistic Latency
A safer architecture would use multiple regenerative relays communicating across independent propagation paths. Consider a network of ten probes, each in different planetary orbits. Rather than relying on a single copy travelling between two nodes, the archive could be encoded with redundancy and transmitted across the network. Each relay would receive incoming transmissions, detect and correct errors where possible, compare independently received versions of the data, reconstruct corrupted fragments, and retransmit verified information toward other relays.
Checksums or cryptographic hashes could verify integrity, while error-correcting or erasure codes could allow the original archive to remain recoverable even when some transmissions are corrupted or some probes fail entirely. These techniques do not guarantee indefinite preservation. Error-correcting codes can reconstruct the original data when corruption remains within their correction capability, but sufficiently severe errors, correlated failures, or the loss of too many relays could make the archive unrecoverable. However, no data storage medium ever invented is guaranteed to retain data with absolute certainty.
The same principle becomes stranger at interstellar distances. If humanity ever becomes an interstellar civilization, relays could be positioned in different star systems. Because the distances between stars are measured in light-years, a transmission could spend years travelling between relays. Considerable amounts of data could be kept in transit by transmitting it back and forth between star systems.
This is due to the simple relation:
The disadvantage is equally significant. Once transmitted, the data cannot be accessed until enough of the archive reaches a relay and becomes reconstructable. An interstellar propagation archive could therefore have access latencies measured in years or decades.
Another disadvantage of this method of data storage is the extreme loss of received signal power across astronomical distances. For an ideal point source, intensity decreases approximately according to the inverse-square law: doubling the distance reduces intensity to one quarter. This means our relay probes would require larger and more complex antennas, precise pointing systems, substantial power, lower transmission rates, or some combination of these requirements.
Distance, normally considered a limitation of communication systems, becomes a component of the storage architecture itself.
Normally, communication systems move information between storage devices. In this architecture, the communication network itself becomes part of the storage system. Information is preserved not by remaining stationary, but by continuously propagating, being reconstructed, corrected, and transmitted again. Whether such a system could ever outperform conventional archival storage is something that would need to be modelled and tested.