Replicated Data Types

A social link based private storage cloud

“…If level N is a fork segment (with more than one nodes, and thus not a single trunk node), and if level N −1 is also a fork segment, then we will show that N = N − 1 must hold, which is impossible. The reason that N must be equal to N − 1 is because either: a) nodes of N would be assigned level N − 1 if the fork of level N − 1 already existed when nodes of level N were created (lines 9-10), or b) the nodes of N would be reassigned to level N − 1 if the fork at N − 1 was created after N was assigned (lines [12][13][14]. Note that, the root node is always a trunk node, as it is created on a single device (no two devices can create an object with the same Oid, as the latter is unique across the devices of the group).…”

“…The algorithm is O(1) in the common cases of adding versions to a version graph with no conflicts (lines 2-5), and when adding one more fork node at the highest-leveled fork segment (lines 9-10). If a fork node is added above that, our algorithm will reassign the level at all nodes below the currently added one (lines [12][13][14], which is more costly. for each V in toUpdate do 14:…”

“…Instead, with this mechanism (inspired by [12]), a content element is guaranteed to be deleted from a system node only when another node takes custody of it. 9) Consistency and privacy properties: The MS , synchronized via our epoch-graph-based synchronization protocol, forms a delta-state Conflict-Free Replicated data type (δ-CRDT, see [13]) and guarantees Strong Eventual Consistency (SEC, [14]). EGSync provides SEC because, provided that all messages are delivered, the MS repository will be identical across all nodes, and a conflict can never occur for the MS data itself; the only potential conflicts in our system are for user files which are presented to user-space consistently by our FS layer.…”

“…Our metadata synchronization algorithm forms a deltastate Conflict-Free Replicated data type (δ-CRDT, see [13]) that achieves Strong Eventual Consistency [14] for our MS . That is, a conflict can never occur for the MS data itself; the only potential conflicts in our system are for user files.…”

Analysis of the Matrix Event Graph Replicated Data Type

“…As the main contribution, we therefore provide an analysis of the degree to which the MEG fulfills consistency, deployability in decentralized scenarios, and scalability: Consistency: Since Matrix provides availability and partition tolerance, in accordance with the CAP theorem [3], the MEG necessarily has to sacrifice strong consistency. We show that the MEG provides Strong Eventual Consistency (SEC) by proving that the MEG is a Conflict-Free Replicated Data Type (CRDT) [4] for causal histories. We compare SEC to Eventual Consistency and Strong Consistency in Subsection IV-A.…”