Step 2: Virtual IBC
In the previous step we added Polymer as a hub in between two Cosmos chains. Now let’s replace one of the chains with a chain that is not natively compatible with IBC, like Ethereum.
Introducing the middle hop alone, does not solve the issue. Even though there’s been some work done on an IBC core implementation in solidity, verifying Tendermint consensus (representing Polymer chain) on the EVM is prohibitively expensive, please refer to this blog post for more background.
The solution provided by Polymer is to outsource the transport (and part of the state) layer to the Polymer chain. This fits within the modular interoperability framework.
The application and transport layer separation have always been part of the IBC design, ever since its conception. This has the benefits of separating the concerns of application and core developers respectively.
Polymer takes this modularity a step further though. Similar to the ideas of modular blockchains introduced by Celestia, Polymer proposes the concept of modular interoperability. This means that all 3 layers of an interoperability protocol (app, transport and state) still form a logical whole, yet they can be decoupled onto different chains.
Virtual transport with vIBC
Instead of running an entire IBC implementation (including an on-chain light client that verifies Tendermint consensus inside the EVM), two simplified components need to be added to Ethereum.
- vIBC core smart contracts (virtual transport in the diagram). These smart contracts will satisfy the required IBC interfaces to handle channel and connection handshakes, and the packet flow (send + callbacks).
The vIBC (virtual IBC) module on Polymer’s side handles the changes related to introducing a virtual transport layer.
- A verifier contract. The verifier contract deployed on the EVM is able to verify the consensus proofs of the Polymer chain. Note that Polymer is also running a client for itself (on behalf of the virtual chain), through the localhost or loopback client.
As consensus proof verification for Tendermint consensus (Polymer chain is a Cosmos SDK chain) is prohibitively expensive, a solution was found to have the verifier contract verify Zero-Knowledge Proofs (ZKP) instead. More on that in the next section on zkMint
Conclusion
We end up with a situation where an IBC application on the virtual chain (e.g. Ethereum) can trigger IBC interactions with the Cosmos chain B, through the vIBC core smart contracts which then outsources the heavy lifting to the Polymer chain running the IBC transport layer on behalf of Ethereum. We call this virtual IBC and Ethereum in this case acts as a virtual chain.