You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
For context, a similar, but more general, idea has already been raised: p2panda/website#202
We need something along these lines for blobs so that a node can be configured to replicate only certain types of blob documents and their pieces, rather than allblob_v1 and blob_piece_v1 documents. To summarise the behaviour we want is:
blob_v1 and blob_piece_v1 are built-in system schema which describe binary data. Nodes are not expected to directly replicate documents following these schema as that would cause a flood of data they are likely not interested in.
a user can define their own schema which contains a field relating to blob_v1 , if a node is configured to support this application schema, then it is expected that they also want all related blob_v1 but not necessarily the blob_piece_v1 documents
a further step the node can take is to specify that, yes, they actually do want to replicate the blob_piece_v1 documents as well
This should be achievable by adding logic to the replication service which includes dependencies of application schema in their target sets when initiating sync requests.
The desired replication behaviour (eg. "I want all blobs and blob pieces for schema X, Y & Z") would need to be expressed in Announce messages as well.
The text was updated successfully, but these errors were encountered:
For context, a similar, but more general, idea has already been raised: p2panda/website#202
We need something along these lines for blobs so that a node can be configured to replicate only certain types of blob documents and their pieces, rather than all
blob_v1
andblob_piece_v1
documents. To summarise the behaviour we want is:blob_v1
andblob_piece_v1
are built-in system schema which describe binary data. Nodes are not expected to directly replicate documents following these schema as that would cause a flood of data they are likely not interested in.blob_v1
, if a node is configured to support this application schema, then it is expected that they also want all relatedblob_v1
but not necessarily theblob_piece_v1
documentsblob_piece_v1
documents as wellThis should be achievable by adding logic to the replication service which includes dependencies of application schema in their target sets when initiating sync requests.
The desired replication behaviour (eg. "I want all blobs and blob pieces for schema X, Y & Z") would need to be expressed in
Announce
messages as well.The text was updated successfully, but these errors were encountered: