This is very cool. Simple enough to reason the invariants easily. I guess one of the key insights is that each data has a canonical server owner which enforces the consistency of the writes of the data at a single place.
I have one question regarding the determination of the latest version of a set of peer data when overlapping transactions occurred.
Suppose initially s1/x=0 at server s1 and s2/y=0 at server s2. Client1 updates s1/x=10 and s2/y=10 at transaction v1, and at the same time client2 updates s1/x=20 and s2/y=20 at v2. Suppose the clients contact the servers in different order and the update messages arrive at the servers in reverse order, such that s1's pending queue is [x.v1=10, x.v2=20] and s2's pending queue is [y.v1=20, y.v2=10].
After client1 and client2 send the make-good command, s1's good queue is [x.v1=10, x.v2=20] and s2's good queue is [y.v2=20, y.v1=10]. When a third client3 tries to read the latest value of x or y, what is the latest value of its peer data? It looks like depending which data client3 starts with, it would get a different version of the peer data? Like starts with s1/x got s1/x.v2 == 20 and s2/y.v2 == 20, while starts with s2/y.v1 == 10 and s1/x.v1 == 10.
Am I missing something or this is the semantic in determining the latest values of peer data?
I have one question regarding the determination of the latest version of a set of peer data when overlapping transactions occurred.
Suppose initially s1/x=0 at server s1 and s2/y=0 at server s2. Client1 updates s1/x=10 and s2/y=10 at transaction v1, and at the same time client2 updates s1/x=20 and s2/y=20 at v2. Suppose the clients contact the servers in different order and the update messages arrive at the servers in reverse order, such that s1's pending queue is [x.v1=10, x.v2=20] and s2's pending queue is [y.v1=20, y.v2=10].
After client1 and client2 send the make-good command, s1's good queue is [x.v1=10, x.v2=20] and s2's good queue is [y.v2=20, y.v1=10]. When a third client3 tries to read the latest value of x or y, what is the latest value of its peer data? It looks like depending which data client3 starts with, it would get a different version of the peer data? Like starts with s1/x got s1/x.v2 == 20 and s2/y.v2 == 20, while starts with s2/y.v1 == 10 and s1/x.v1 == 10.
Am I missing something or this is the semantic in determining the latest values of peer data?