Create op-supervisor.mdx

pages/stack/interop/_meta.json

@@ -2,5 +2,6 @@
     "explainer": "Interop explainer",
     "cross-chain-message": "Anatomy of cross-chain message",
     "supersim": "Supersim Multichain Development Environment",
-    "superchain-erc20": "SuperchainERC20"
+    "superchain-erc20": "SuperchainERC20",
+    "op-supervisor": "OP Supervisor"
 }

pages/stack/interop/op-supervisor.mdx

@@ -0,0 +1,73 @@
+---
+title: OP Supervisor
+lang: en-US
+description: Learn the basics of OP-Supervisor.
+---
+
+import { Callout, Tabs, Steps } from 'nextra/components'
+
+# OP Supervisor
+
+<Callout>
+  Interop is currently in active development and not yet ready for production use. The information provided here may change. Check back regularly for the most up-to-date information.
+</Callout>
+
+OP Supervisor is a critical service designed to verify cross-chain messages and manage the safety of interoperability between different chains in the OP Stack. It acts as a hub for indexing data that every `op-node` needs to cross-verify with other chains.
+Some features and benefits include:
+
+*   Enables secure cross-chain message passing on the Superchain
+*   Provides a centralized point for managing interoperability data
+*   Supports multiple networks simultaneously
+*   Offers potential for public endpoints to aid in node synchronization
+
+## Cross-chain message verification
+
+Verifies messages between different chains in the OP Stack, reducing the risk of invalid or malicious cross-chain interactions.⁠ It centralizes the verification process, which reduces the complexity of operating individual nodes⁠.⁠
+
+*   `op-geth`: queries `op-supervisor` during block-building to verify if a message is sufficiently safe to include. This process involves checking each executing message and potentially undoing transactions if conflicts or unknown states are encountered.
+*   `op-node`: queries cross-chain safety information and coordinates safety updates between OP stack nodes and `op-supervisor`. It uses the API provided by `op-supervisor` to perform actions like:
+    *   Updating and retrieving various safety levels
+    *   Checking and returning the `cross-unsafe` head for a given chain
+    *   Attempting to promote a block to `cross-safe` status
+    *   Attempting to finalize an L2 block based on L1 finality
+
+## Log data indexing and management
+
+Acts as a hub for indexing data that every `op-node` needs to cross-verify with other chains, centralizing the process of managing interoperability data. Maintains the integrity of cross-chain interactions by tracking safety changes⁠ across the Superchain, ensuring consistent application of invalid dependency resolutions. ⁠
+
+`op-supervisor` indexes two types of cross-chain dependencies:
+
+*   Interop messages (events): `op-supervisor` maintains an events-index per L2 chain, which determines message-dependencies to check if blocks are safe
+*   L1 DA (data availability): `op-supervisor` tracks the L1 DA for L2 blocks and maintains a DA safety-index per L2 chain, which helps determine how to rewind L2 chains to resolve invalid dependencies
+
+## API for cross-chain safety
+
+Provides an interface for `op-node` to query cross-chain safety information and coordinate safety updates between OP stack nodes and `op-supervisor⁠⁠`. OP-Supervisor uses a global read API to determine **message safety** and **block safety,** utilizing both the events index and the safety index (See op-supervisor's [safe data indexing](#safe-data-indexing-and-management)). The API is designed to handle potential L1 reorgs that can affect the validity of cross-chain messages⁠.
+
+Key API methods include:
+
+| Method                                    | Description                                                                           |
+| ----------------------------------------- | ------------------------------------------------------------------------------------- |
+| `UnsafeView` and `SafeView`               | Returns the Local and Cross heads for their respective levels                         |
+| `DerivedFrom`                             | OP Nodes use to check the L1 source of the Supervisor (needed for Safe Head tracking) |
+| `UpdateLocalSafe` and `UpdateLocalUnsafe` | Tells the Supervisor when the Node's heads change                                     |
+| `Finalized`                               | Returns the Finalized Head                                                            |
+| `UpdateFinalizedL1`                       | Signals to the Supervisor new finality signals                                        |
+| `CheckMessage`                            | Checks logs in the DB directly in tests, can also use `CheckMessages` in geth         |
+
+For a full listing of API names, see the [`op-supervisor` client](https://github.com/ethereum-optimism/optimism/blob/develop/op-service/sources/supervisor_client.go).
+
+## RPC access to all chains
+
+The `op-supervisor` requires RPC access to all chains in the dependency set. This allows `op-supervisor` to verify cross-chain messages and sync data across multiple networks simultaneously, such as OP Mainnet and Base nodes using the same instance.
+
+Benefits of this approach:
+
+*   Cost-effective: Running `op-supervisor` is cheaper than running a full node for each chain in the dependency set.
+*   Scalability: As the number of chains in the Superchain grows, `op-supervisor` can handle the increasing complexity of cross-chain interactions.
+*   Improved reliability: It enables a more redundant setup, which is good for stability in the growing ecosystem.
+
+## Next steps
+
+*   Want to learn more? Read our guide on the anatomy of a [cross-chain message](/stack/interop/cross-chain-message) or check out this [interop design video walk-thru](https://www.youtube.com/watch?v=FKc5RgjtGes).
+*   For more info about how OP Stack interoperability works under the hood, [check out the specs](https://specs.optimism.io/interop/overview.html).