Closed kdenhartog closed 4 years ago
Just a thought:
Layer 1: Message Layer - required encryption, replication, and routing bits that must be integrated into messages
Layer 2: Storage Layer - messages are stored, managed, and permissioned as a coherent singular mass
Layer 3: Query Layer - how the storage layer groups, indexes, and exposes the data in the storage layer
Just a thought:
Layer 1: Message Layer - required encryption, replication, and routing bits that must be integrated into messages
Layer 2: Storage Layer - messages are stored, managed, and permissioned as a coherent singular mass
Layer 3: Query Layer - how the storage layer groups, indexes, and exposes the data in the storage layer
I'm good with those names.
I suggest we delay this layering exercise until after we name the key roles. In particular, I would wait for [w3c/did-use-cases] Relying Party -> Requesting Party (#84)
-1, Messaging at the base layer is concerning, maybe? Needs more discussion, I'd like to do this at the end ... when we have a full idea of what all of the layers are.
@msporny by messaging, I simply mean: the actual atomic objects sent between SDS instances, and the definitions of their required/standard properties and values the SDS system needs to properly handle them. Could also be referred to as Object Layer.
Could also be referred to as Object Layer.
Right, agreed that "Object Layer" might be better... Getting a full list of features might be more helpful... once we have all of the features identified, then we can sort them into layers, IMHO. Once we sort them into layers, we can name the layers.
Another proposal here: https://github.com/decentralized-identity/secure-data-store/issues/74
In an effort to consolidate the threads around this discussion, I'll repost the proposals from this thread there and close it. See this https://github.com/decentralized-identity/secure-data-store/issues/74#issuecomment-645676378 for more details
I recognize this is a bikeshed issue, so we should be mindful of this. Especially as we're still aligning on a mental model and architecture. To get things started though, I propose this set of names for the layers:
Layer 1: Storage layer Layer 2: Replication layer Layer 3: Integrity layer