HackMDCSM LTI considerations v2
This is a revised version of the original doc
TL;DR
Implementing a CSM long-term incentive program is proposed by awarding well-performing Node Operators with a reduced bond for the very first validator (same benefit as for the Early Adoption). The CSM Performance Oracle will maintain performance assessment and benefits application.
Full version
LTI - long-term incentive. "A long-term incentive, as the name suggests, is a vehicle that has an extended time horizon (generally greater than one year) and that can be a strategic compensation vehicle to promote long-term retention and alignment with company goals."
– Mercer
For CSM Node Operators, LTI can be treated as a tool to incentivize their performance over a long time.
Long-term operation within CSM is beneficial for Lido in terms of decentralization. Note that having ten small Node Operators is more desirable than having one large Operator maintaining the same number of validators. This fact should be taken into account when deciding on the actual benefits.
An incentive program is proposed to make long-term performance beneficial and attractive for the Node Operators. This program should add a small yet meaningful advantage for the Node Operators performing well over the long timespan in CSM.
To be eligible for the program benefits, the Node Operator must perform above the performance threshold for longer than a year and acquire no penalties during the period.
There are two main constraints for the CSM Node Operators set within CSM: staking fee and bond size.
With the current architecture of the Staking Router, variable staking fees can be implemented only with the help of "tricks" like full fee allocation to the module and further return of the Lido DAO treasury commission back; let's not consider this option.
What we have left is bond size. Given the CSM bond curve mechanism, bond requirements can be adjusted for any validator controlled by the Node Operator (ex., reduced bond for the very first validator or each fifth validator). Hence, let's focus on this option.
Another important consideration is the CSM Early Adoption program, which offers beneficial conditions and early access for proven solo-stakers and active testnet participants. Both categories are selected based on the performance data available for them. In this case, the incentive program is no different since it aims to benefit the well-performers. Hence, it is reasonable to offer CSM well-performers the same benefit as the Early Adopters - reduced bond for the very first validator.
CSM accounting is built so that any changes in the bond curve will have an immediate effect, and the required bond will be recalculated. It means that after applying the beneficial curve, exiting Node Operators will have excess bonds that can be claimed or used to create more validators.
Operational costs of the incentive program should also be accounted for. Thus, the best way is to automate the benefits application or permissionlessly initiated by the Node Operators. There are two options to achieve that:
- CSM Performance Oracle can assess and apply benefits for the Node Operators;
- CSM Performance Oracle can explicitly report the poor performance to CSM (on-chain). MEV stealing and initial slashing penalties application flag is recorded on-chain. A permissionless method checks if enough time has passed since the Node Operator creation and if any penalties or poor performance are registered. If all conditions are met, benefits are applied;
While the second approach might look more appealing regarding trust assumptions, it is way more complicated and less flexible than the first one. Hence, it is proposed that we proceed with the first option.
The following stuff will be required for implementation:
- Update CSM Performance Oracle code;
- Grant CSM Oracle contract with the permission to set beneficial curve for the Node Operators;
