meeting minutes: 2020-11-09

Author: Wollac

meeting-minutes/2020-11-09.md

@@ -0,0 +1,55 @@
+# Chrysalis RFC Call - 2020-11-09
+
+## Participants
+
+- Wolfgang
+- Andrea
+- Luca
+- Gal
+- Thibault
+
+## Topics
+
+### RFC Status
+
+#### Merged
+
+https://iotaledger.github.io/protocol-rfcs/
+
+- [RFC-0005](https://github.com/iotaledger/protocol-rfcs/blob/master/text/0005-white-flag/0005-white-flag.md) (White Flag):
+  - Remove old Bundle/Tx structure:
+    - done: https://github.com/iotaledger/protocol-rfcs/pull/30
+- [RFC-0008](https://github.com/iotaledger/protocol-rfcs/blob/master/text/0008-weighted-uniform-random-tip-selection/0008-weighted-uniform-random-tip-selection.md) (Weighted URTS):
+    - Everyone is happy with the changes to reflect the current implementation (in Hornet)
+    - @all approve
+- [RFC-0012](https://github.com/iotaledger/protocol-rfcs/blob/master/text/0012-milestone-merkle-validation/0012-milestone-merkle-validation.md) (Milestone Validation):
+  - Remove old Bundle/Tx structure:
+    - State mutations -> UTXO mutations (circular transactions are not possible with UTXO)
+    - Tx Hash -> Message ID
+    - Blake2b-512 -> Blake2b-256
+  - @Wolfgang create PR
+
+### Open Questions
+
+#### `Network ID` field in the message
+
+- Purpose:
+    - Distinguish between different (private) networks
+    - Detect/handle breaking protocol changes independently of the network/peering layer
+    - Prevent replays of the same message in different networks without PoW
+    - Have network/version as part of the Message ID (for MSs, queries, ...)
+    - Since the MS payload signs the Message IDs, replays of the MS payload to a different network is not possible.
+- Decision:
+    - It should be included as a single message field.
+    - It should encode the network (mainnet, testnet, private x, ...) as well as the protocol version.
+    - It should be a `uint64`. This offers a good compromise between size and accidentally having the same value for incompatible situations.
+    - Local snapshots (and probably peering) should use the same value.
+    - It should not be signed to potentially allow replays of the same payload (but with different Message ID and new PoW).
+    - Nodes should reject message with an incompatible `Network ID`. (What actually causes a protocol version bump is a case-by-case definition, but as a rule of thumb consensus changes are breaking.)
+    - (Perma)nodes trying to sync across different protocol versions is a special case and does not need to be considered for regular nodes (same premise as when introducing white-flag).
+    - How the `Network ID` is constructed is an implementation decision. It does not need to be part of the Message RFC, but it is still very relevant for the node implementations. Options:
+        - Hash(`Network Name`+`Protocol Version`) (probably the most promising)
+        - Use a couple of static bytes for the "Network"
+        - Combination of the above
+- @Gal update the PR to include those changes
+- After that, [RFC-0025 (Local Snapshots)](https://github.com/iotaledger/protocol-rfcs/pull/25) needs to be adapted