The Cosmos Hub v9 upgrade went live on March 15th. This upgrade added the code for Replicated Security to the Cosmos Hub. Due to our renewed focus on validator outreach and communications, v9 went very smoothly, with minimal downtime.
Replicated Security is now enabled on the Cosmos Hub, although it won’t really be used until the first consumer chain launches.
We’ve been supporting the Neutron team as they move towards launch as the first RS consumer chain. They’ve had their proposal on the forum for a few weeks, and it will go on chain soon. Our team has been double checking their implementation and the parameters to be used in their on chain governance proposal. We’ve also been assisting them with thinking through the question of how to minimize costs for smaller validators. One solution we’ve been considering is a feature called soft opt-out. With soft opt-out, a percentage of the smallest validators can opt out of validating a consumer chain without affecting the security of the consumer chain (although it might have slightly higher downtime during upgrades). Neutron may use this to allow the smallest 5% of validators to opt out. Given the distribution of the Cosmos Hub validator set, this is actually around 75 validators.
We’ve been working on some documentation for Replicated Security. They are fairly complete, but could probably still use some improvement. Let us know with an issue if anything is unclear.
We’ve been helping the Stride team think through the best way to bring secure and user-friendly liquid staking to the Cosmos Hub. One promising approach is to add an additional limit to the existing liquid staking module, which would enforce a global cap on the percentage of Atoms that can be liquid staked. This would also apply to liquid staking done through Interchain Accounts, which is currently not limited at all.
This global limit would make the adoption of liquid staking more secure by imposing this limit, while also improving the user experience of liquid staking by making it seamless to adopt.
Opt-in Security is a form of Interchain Security where validators can choose which consumer chains they validate on, as opposed to Replicated Security, which uses the full validator set of the Cosmos Hub for each consumer chain. We’ve sketched out a rough design for a first version of opt-in security here.
The Stride team has been working on a first version of an implementation of our protocol enabling Cosmos chains which are currently running as standalone chains to switch to Replicated Security. We’re currently working on bringing this into mainline Interchain Security, working through all possible edge cases, and testing it in our test suite.
Work has been progressing quickly on code which will allow the Cosmos Hub to detect and slash validators for light client attacks. This is a key component of our move towards an trustless consumer chain model, where faulty or malicious consumer chains cannot cause any disruption to the Cosmos Hub. Once this work is complete, we will be able to remove the governance-gating of slashing, instead allowing misbehaving validators to be immediately identified, cryptographically verified, and slashed without an extra confirmation from governance,
We’re also working on code that will allow the Cosmos Hub to detect and verify that a validator has experienced excessive downtime on a consumer chain. Once equivocation verification and downtime verification are complete, the Cosmos Hub will be able to verify all misbehavior of validators on consumer chains, instead of relying on the consumer chains themselves to support it. This will make Replicated Security simpler and more robust.