docs(arc-5): ITS hub express

arc-4.md

@@ -0,0 +1,91 @@
+
+# ARC-4: ITS Hub Express
+## Metadata
+
+
+
+-  **ARC ID**: 4
+
+-  **Author(s)**: CJ Cobb
+
+-  **Status**: Draft
+
+-  **Created**: 2025-03-20
+
+-  **Last Updated**: 2025-03-20
+
+-  **Target Implementation**: Q2 2025
+
+
+## Background
+
+GMP Express is a feature that allows GMP calls to be executed at the destination contract prior to the GMP call
+being fully verified by Axelar. This is most commonly used for token transfers. The express relayer will facilitate
+the transfer using its own tokens, and will be reimbursed those tokens later when the verified GMP message arrives.
+
+This process uses the message id to map an expressed GMP call to the actual GMP call. However, transfers using ITS
+hub actually involve two GMP calls (source to hub, and then hub to destination), each with their own message id.
+Therefore, when the GMP message is initially sent at the source, the final message id that will be seen by the
+destination is not yet known. Thus, the existing design of GMP express does not work with ITS hub.
+
+## Proposed Solution
+
+The proposed solution is to express the first GMP call of the transfer, specifically the GMP call from the source
+chain to the hub. This is typically much slower than the second hop, since fully verifying the first hop requires
+waiting for finality. Expressing to the hub also simplifies operational overhead for express relayers, since the
+relayers only need to maintain token balances on one chain (Axelarnet), instead of many destination chains.
+
+### Flow
+
+```mermaid
+flowchart LR
+    User
+    ExpressRelayer["Express Relayer"]
+    Edge["ITS Edge Contract"]
+    Gateway
+    Amplifier 
+    Gas["Gas Service"]
+    Hub["ITS Hub"]
+    GMPRelayer["GMP Relayer"]
+
+    subgraph Ethereum
+    User --InterchainTransfer--> Edge
+    Edge --CallContract--> Gateway
+    Edge --Pay gas for express call--> Gas
+    end
+    Gas -.observe express gas paid.-> ExpressRelayer
+    ExpressRelayer --express execute (with tokens)--> Hub
+    Gateway -.observe gmp call.-> GMPRelayer
+    GMPRelayer --verify msg--> Amplifier
+    subgraph Axelar
+    Amplifier --deliver verified msg-->Hub
+    end
+    Hub --refund original tokens--> ExpressRelayer
+```
+
+TODO: deployment flow
+
+TODO: sequence diagram
+
+### Considerations
+
+Express execution of arbitrary contract calls is not universally safe, and usually apps opt in to express execution
+of calldata via defining a particular endpoint on their contract. However, when we express to ITS hub instead of to
+the receiving app, apps cannot opt in or out of call execution in the normal way. We need to design a way to allow
+apps to opt in.
+
+Two options:
+* Maintain a whitelist in ITS hub, which can be updated by governance.
+* Set a flag in the ITS packet sent to the receiving app signaling whether the call was expressed. Apps can choose
+to ignore expressed calls.
+
+The second approach requires changing the ITS packet format, which would require upgrading existing ITS edge
+contracts, and would be an additional requirement for integrators. (would this also break existing ITS apps?).
+
+
+### Components to build
+
+* Express execute endpoint and refund logic on ITS hub
+* ITS edge contract on Axelar (for minting)
+* Enable token minting on Axelar in core
+* Functionality for apps to allow attaching calldata to express execution