It is expected that message histories older than 30 days (or some other number of days) won't be frequently accessed by the users. Therefore, old historical messages can be safely migrated to another file-sharing platform in order to save storage for the store nodes. As the result, store nodes will only maintain a pointer to the migrated data that can be given to the querying nodes to retrieve them from the file-sharing layer.
When tackling this problem, there are few things to consider
how to group messages for migration
how to derive a pointer to the migrated group of messages, also depends on which file-sharing system we are going to use and how data is addressable in that system.
Solution ideas
Assuming that the file-sharing system is key-value-based storage, then the Merkle tree root of a block of messages is one way to get the block pointer i.e., which constitutes the key component of the message block in that file-sharing system.
Problem definition
It is expected that message histories older than 30 days (or some other number of days) won't be frequently accessed by the users. Therefore, old historical messages can be safely migrated to another file-sharing platform in order to save storage for the store nodes. As the result, store nodes will only maintain a pointer to the migrated data that can be given to the querying nodes to retrieve them from the file-sharing layer.
When tackling this problem, there are few things to consider
Solution ideas
Assuming that the file-sharing system is key-value-based storage, then the Merkle tree root of a block of messages is one way to get the block pointer i.e., which constitutes the key component of the message block in that file-sharing system.