What does synchronisation mean in cloud storage?¶

Synchronisation is the process by which a cloud client detects a change in one location and propagates it to the provider and other connected locations. It can make a file appear, change or disappear on a device without a person performing that same action there.

A change moves through stages¶

When a file is saved in a synced folder, the client normally:

1. detects the local change;
2. identifies the account, folder and provider object;
3. uploads content or a change request;
4. receives confirmation and updates its local sync state; and
5. makes the provider change available to other devices and users.

The reverse route is also possible. A provider-side edit or another device's upload can be detected and downloaded by the local client. “Synchronised” therefore describes a relationship between systems, not a direction of travel.

The original event and propagated events are different¶

Dave saves briefing.docx on Surface Laptop DEV-A at 18:04. Its OneDrive client submits the change; the provider accepts object OBJ-7712, version V-8, at 18:05. A second laptop receives the full copy at 18:07, while Dave's Pixel updates only its object listing at 18:12.

The creation time on Laptop B records arrival there, not authorship of the document. The Pixel can know the object exists without holding the full file. The next useful records are DEV-A's application/filesystem activity, the provider acceptance event and each receiving client's mapping and transfer state.

Synchronisation is not necessarily immediate or complete¶

A device may be offline, paused or configured for selected folders. Uploads can fail, and two devices can edit the same version before learning about each other's changes. The service may merge the changes, choose one version or create a conflict copy.

The available records may include:

• provider object, version and activity events;
• client device or installation identifiers;
• sync databases linking cloud IDs to local paths;
• queued, completed and failed transfer entries; and
• local filesystem and application activity.

Retention and detail vary, so an apparent gap in one source should be compared with the other parts of the sequence.

Attribute the cause, not every consequence¶

Automatic propagation can explain why matching content appears on several devices or why a deletion spreads. It does not automatically explain who caused the first change. Conversely, later opening, editing or sharing on a receiving device may show deliberate interaction after an automatic arrival.