Data Origin
The concept of the Data Origin is important within Panzura's hybrid cloud file platform CloudFS. When a file is created, the node on which it was created is designated the Origin, and this is recorded in its metadata. The origin always knows which node currently holds the file or range lock, regardless of whether the file is currently locked for editing.
Because each node within a CloudFS deployment has a complete, up-to-date copy of the metadata of the entire global file system, every node understands which node to ask, when attempting to take over the lock. It makes this request peer-to-peer, allowing for a real-time exchange of lock ownership if granted, as well as any data that has changed within the last minute.
If the lock is sitting with the Origin, it will either approve the request, or deny it if the file or range is open for editing. If the Origin does not currently hold the lock, it tells the requesting node which location to request it from.
Because each node within a CloudFS deployment has a complete, up-to-date copy of the metadata of the entire global file system, every node understands which node to ask, when attempting to take over the lock. It makes this request peer-to-peer, allowing for a real-time exchange of lock ownership if granted, as well as any data that has changed within the last minute.
If the lock is sitting with the Origin, it will either approve the request, or deny it if the file or range is open for editing. If the Origin does not currently hold the lock, it tells the requesting node which location to request it from.