JIP-8: Adopt TipRouter NCN (Protocol Development)

JIP - Official
1

Abstract

This JIP will authorize the Jito DAO to execute a series of transactions, assume a number of new powers, and signal approval for actions associated with the Jito DAO adopting a deployed instance of the open-source TipRouter NCN code.

The TipRouter NCN will assume programmatic operations related to the distribution of MEV tips generated from the Jito Tip Distribution protocol every epoch, and, as part of these programmatic operations, will route 3% of all MEV tips collected to the DAO treasury and the NCN network participants.

This proposal will also have a number of ancillary effects, including:

  1. Signaling approval for changes to the parameters of StakeNet.
  2. Executing a reallocation of a large portion of the DAO treasury to support TipRouter NCN operations.
  3. Granting the DAO upgrade authority over the TipRouter NCN code.
  4. Granting the DAO blacklist authority over vault managers
  5. Granting the DAO modification authority over the fee take rate for the TipRouter NCN.

Motivation

In simple terms, this proposal will achieve two key effects:

  1. Create a new source of fees for the DAO from the MEV tips generated from the Jito TipRouter protocol (contingent upon the passing of JIPs 8 and 10, 2.7% of which will be routed to the Jito DAO Treasury, with an additional 0.30% going to NCN network participants).
  2. Decentralize and harden the process of distributing MEV tips by placing the process under the control of the DAO, and ensuring trustless operations via multiparty consensus and economic security.

These two effects represent a major step in the broader decentralization and sustainability of the Jito Network, while simultaneously bringing one of the Jito Network’s most powerful features – the processing of MEV tips – under the control and oversight of the DAO.

Key Terms

  • MEV (Maximal Extractable Value) Tips: Tips paid by users participating in a blockspace auction powered by Jito-Solana.

  • TipRouter NCN (Node Consensus Network): A decentralized network that takes responsibility for programmatically handling the distribution of tips collected by the Tip Distribution Protocol.

  • Jito DAO: A decentralized autonomous organization responsible for overseeing the governance and operations of the Jito Network.

  • Tip Distribution Protocol: A mechanism that aggregates SOL (the native cryptocurrency of Solana), generates a merkle tree and merkle root for distribution, and transfers SOL to the appropriate users/accounts.

  • Merkle Root: A cryptographic representation of a set of data (in this case, validator tips) that is used to verify the integrity of data stored or transmitted in the network.

  • StakeNet: An on-chain program responsible for managing the stake pool for JitoSOL.

  • NCN: A Node Consensus Network. A network that uses Jito (Re)staking to come to consensus regarding an on or offchain process.

  • Vault Restaking Tokens (VRTs): Tokens representing staked assets in the NCN vaults, which provide economic rewards to participants.

  • DAO Treasury: The accumulated rewards from various fee streams generated from the Jito Network and directed to the DAO.

  • Upgrade Authority: The ability to modify or upgrade the TipRouter NCN. Initially held with parallel authority by the Jito DAO Security Council and the Jito DAO, with the potential for migration to solely DAO governance.

  • Node Operator: An entity tasked with managing hardware or software infrastructure requirements for NCNs.

  • Vault manager: An entity tasked with managing the delegations to NCNs, distributing NCN rewards, and creating VRTs.

Specifications

Today

The Jito Tip Distribution Protocol’s mechanism currently functions in the following manner:

  • Each validator has a unique tip distribution account, owned by the Jito Tip Distribution program, where SOL is collected for a given epoch.
  • At the end of the epoch, an off-chain process takes place which produces a merkle tree and merkle root for each validator’s tip distribution account.
    • Validators can run this process themselves or delegate it to another party.
    • The merkle tree is intended to distribute any SOL in the account to the validator based on their MEV commission rate. Remaining funds are sent to stakers’ stake accounts pro-rata.
  • After the merkle root has been uploaded, merkle proofs are uploaded on-chain and the program will transfer SOL from the tip distribution account to the validator’s vote account and stake accounts.
    • For each claim, a ClaimStatus account is created to avoid double spends.
    • ClaimStatus accounts are stored on-chain for 10 epochs. After 10 epochs, the ClaimStatus accounts can be closed and SOL is refunded to the account which paid the rent-exempt SOL.

Today, merkle root upload authority is permissioned with opt-in from each Jito-Solana validator. This mechanism creates a single point of potential failure with adverse impacts on network stakeholders and lacks transparency on the calculation.

TipRouter NCN

The TipRouter NCN will leverage Jito (Re)staking to empower a set of node operators to come to consensus on the correct merkle root for tip distribution.

The mechanism of the TipRouter NCN is detailed below:

  • Validators can delegate the merkle root upload authority to a program derived address owned by the NCN, giving it permission to upload merkle roots.
  • After epoch rollover, node operators will each compute the merkle tree and merkle root for each validator and upload it on-chain.
  • On a periodic basis, consensus of node operators will be checked. After ⅔ of stake agrees on a merkle root for a given validator, a cross program invocation will take place to upload the merkle root to the validator’s tip distribution account.

All validators leveraging the Jito Tip Distribution protocol with the merkle root upload authority GZctHpWXmsZC1YHACTGGcHhYxjdRqQvTpYkb9LMvxDib will be opted into the TipRouter NCN automatically. Transitioning from the current MEV tip distribution system to the TipRouter NCN as described above hinges on a number of factors for the DAO to consider and actions that need to be taken.

Economic Assumptions/NCN Parameters

This proposal suggests that the TipRouter NCN launches with a 3% fee on all MEV tips distributed. Of that 3%, 2.7% will be routed to the DAO in the form of JitoSOL, while the remaining .30% will be routed to NCN participants, including node operators, vault managers, and restakers, per JIPs 8 and 10.

The proposed parameters for TipRouter NCN fees were made using the following assumptions:

For the month prior to Nov 20th, the annualized run-rate for Solana MEV tips was 9.8mm SOL, or ~$2.3 billion, assuming $236 SOL. With 157 epochs per year, this implies average economic value at risk of $14.7mm per epoch.

418×168 4.39 KB

Using $2.3billion in tip fees per year in dollar terms as a baseline from recent data, assuming $650mm in total JitoSOL and JTO deposits in the TipRouter NCN ensures that there is approximately 45x the amount of economic security at stake as there are MEV tips distributed in a given epoch:

777×364 9.97 KB

At $500mm in NCN deposits restakers will generate a rewards rate of .68% in programmatic fees with the above assumptions. The following table is provided for the sake of economic modeling:

797×432 11.6 KB

Below shows the fees routed to the DAO Treasury using the above assumptions:

792×434 11.8 KB

Full calculations can be found here.

Deposit Assets

The TipRouter NCN will accept any Solana LST as a source of economic security collateral. However, JitoSOL deposits will be favored by weighting rewards towards vaults depending on their percentage composition of JitoSOL. JitoSOL earns 2x the rewards vs. other LSTs. For example, a vault with 80% JitoSOL will receive 1.8x the rewards compared to a vault comprised of 0% JitoSOL.

These admin rights to adjust these weights will initially be managed by the Jito DAO.

Contingent upon the passing of JIP-10, JTO will also be accepted as a collateral asset.

DAO assets

The capital requirement to cover rent for 10 epochs of ClaimStatus accounts is significant. As of October 14, 2024, it was 17,800 SOL, or $2.7M USD at $150 SOL.

In order to ensure the TipRouter NCN can meet this requirement, the DAO is being asked to transfer 25,000 of its SOL to the keeper program. The program will take custody of these funds and use it to fund the rent-exempt amount to hold 10 epochs worth of ClaimStatus accounts on-chain. These funds can be reclaimed by the DAO in the future.

StakeNet parameters

Validators are theoretically capable of designating an alternative merkle root upload authority, and in doing so bypass the NCN’s fees on MEV tips. An untrusted merkle root risks serious financial loss for JitoSOL holders and network participants generally. This would enable accidental or malicious modification to the MEV tip distribution, including theft of all MEV rewards generated by validators using an untrusted merkle authority. MEV rewards represent up to 20% of total staking rewards currently. The potential loss poses great risk to the yield and reputation of JitoSOL and Jito-Solana.

In order to establish guardrails against this behavior, the passing of this proposal will signal approval for modifications to StakeNet such that any validators that use an unapproved merkle root upload authority after some activation epoch are immediately unstaked.

Specifically, in the StakeNet Steward program, after the passing of this proposal there shall be a field added to the configuration: ‘approved merkle root upload authority.’ At some future epoch activation date, ‘compute instant unstake’ will require the validator history merkle root upload authority to match the configured merkle root upload authority for the TipRouter NCN. If a validator ever uses a different merkle root upload authority after the activation epoch, they are immediately unstaked.

Upgrade permissions

Upon the passing of this proposal, upgrade authority for the TipRouter NCN will be managed by the Jito DAO, with parallel authority from the Jito DAO Developer Council. This ensures superior response time in the event of a bug, hack, or other flaw with the instance of the open-source TipRouter NCN code.

Fee and Vault Manager inclusion parameters

All vault managers will be accepted by the TipRouter NCN by default, pending a review from the Developer Council and/or the DAO regarding the quantity of assets their vault will be adding to the NCN. Minimum SOL to be added as a vault manager will initially be set to 5,000 SOL-equivalent units, and minimum JTO to be added to the vault is $500,000 equivalent.

These minimums and the asset review process have been put in place to mitigate DOS attacks in which a malicious party creates a series of empty vaults and maxes out the NCN’s parameters, which are currently set to 64 vault managers and 256 node operators. Additionally, vaults which source multiple LSTs as collateral will require periodic pricing updates, to be managed by the Developer Council or the DAO.

Node operator set will be determined by vault managers, with either the DAO or the developer council responsible for adding them to the operator set. This will leave the functional node operator set – meaning, TipRouter NCN node operators that actually receive stake – entirely in the hands of the vault managers, which will give any node operators the opportunity to participate in the TipRouter NCN, contingent on their ability to successfully petition one of the vault managers. This minimizes the scope of ongoing governance obligations. Vault managers specialize in NO assessment and risk management. More specific restrictions can be implemented by the DAO if the NO performance is below expectations.

The DAO will maintain blacklist authority over the vault manager set, however. One possible edge case is a hostile vault managers attracting enough economic security to take control of the network, and blacklist authority helps to mitigate this risk.

Ideally TipRouter will transition to fully trustless inclusion of vault managers and node operators over time, but shared authority between the DAO and the Developer Council over certain inclusion parameters is necessary to mitigate edge case attacks in the short to intermediate term.

The DAO will additionally have upgrade authority over the TipRouter NCN fee take rate, initially set to 3%.

Benefits/Risks

Benefits:

  • New source of fees for the DAO

  • Decentralizing MEV revenue distribution

  • Bringing new powers and responsibilities under the purview of the DAO

Risks:

  • NCN smart contract risk

  • Fee increase on MEV tips could alter Solana economic dynamics

  • Third-party risk associated with node operators and vault managers

Outcomes

The passing of this proposal will:

  1. Transfer the responsibility of MEV fee distribution to a deployed instance of the TipRouter NCN code
  2. Signal approval for changes to the parameters of StakeNet
  3. Execute a reallocation of a large portion of the DAO treasury to support TipRouter NCN operations
  4. Grant the Jito DAO Developer Council and the Jito DAO parallel upgrade authority over the TipRouter NCN code.
  5. Grant the Jito DAO Security Council and the Jito DAO blacklist authority over the vault manager set
  6. Transfer to the DAO modification authority over the fee take rate for the TipRouter NCN

Cost Summary

This proposal will incur no direct cost to the DAO. However, there will be opportunity cost in the form of using treasury JitoSOL, as described above.

EDIT: This proposal was edited on 11/30 with the following major changes:

1. The original proposal included a deposits cap for the TipRouter NCN. This proposal will include no such cap.
2. This proposal includes JitoSOL deposit weights and places the parameters for the weights under the control of the DAO.
3. The original proposal placed VRT manager set powers under the control of the DAO. This proposal reflects a modification to the TipRouter code such that all VRTs are accepted by default.
4. This proposal includes the ability for the DAO to blacklist certain vault managers.
5. This proposal includes updated economic assumptions with more recent data.

EDIT: This proposal was edited on 12/17 to include additional information about shared DAO/Developer Council authorities related to the addition of vault managers and node operators, as well as to include information about the maximum number of vault managers and node operators in the initial Tiprouter parameters.

12 Likes
2

Concurrently to and contingent upon the passing and implementation of this proposal, Jito Labs will reduce our block builder fee from 5% to 3%. This reflects the lessened infrastructure requirements and responsibilities placed on Jito Labs as a result of this proposal, but ensures Jito Labs remains incentivized to continue covering the costs associated with the upkeep and ongoing development of the Jito Labs Block Engine.

Between the new fees introduced by JIP-8 and the fee reduction from Jito Labs, the aggregate fee placed on MEV tips will rise from 5% to 6% after the passing and implementation of this proposal.

Jito Labs supports JIP-8 and believes it aligns with the broader mission of democratizing MEV and creating an open, decentralized financial system on Solana.

7 Likes
3

I would like to see a corresponding increase in funding and commitment for StakeNet to be more timely with end-of-epoch calculations.

4

Thanks for this proposal, this is an enormous change with a lot of details to unpack.

I appreciate the tables showing some of the ratios on the economics side, what I’d like to understand more fully is why the NCN fee was chosen at 3% and why 1.13% is considered an appropriate NCN rewards rate?

Further, why is the NCN/DAO split 5/95% vs 10/90 or 20/80 etc? The NCN assumes the operational risk and costs while the DAO per this proposal has no exposure aside from the opportunity cost of the 25k SOL rewards? The fee could also be designed to first establish a separate state bond fund to replace the DAO’s 25k SOL, whereafter a normal DAO/NCN split is established.

With regards to the approved merkle root authority, I would like to see a pathway for future alternative merkle root authorities to be established and approved.

Regarding “Initial deposits to TipRouter NCN will be in the form of SOL LSTs accepted by VRT partners, including JitoSOL. Any change to accepted assets is subject to DAO governance.”

  • It is unclear how VRT partners can specify acceptance of LSTs while the DAO can also have ability to change accepted assets, this appears to be a conflict of authority.
3 Likes
5

Re7 Labs would like to express interest in being added to the vault manager set for this proposal.

We will shortly be launching an Re7SOL restaking product that leverages Jito’s underlying infrastructure and focuses on jitoSOL as a deposit asset. Further, this will have a differentiated product by nature of being an early restaking pool using Renzo integration with Jito. We believe our addition to the vault manager set will be a continued value addition for Jito and for other partners in the ecosystem like Renzo.

We currently run Re7LRT and a number of other restaking vaults within the EVM ecosystem, and are excited about the opportunity to expand our restaking vaults to Solana by leveraging Jito.

Re7 has been providing liquidity in DeFi since 2019 having deployed over $300m of assets. We bring practical experience to risk management from years of managing stablecoin and ETH yield strategies as well as further strategies like our Liquid Token fund. As DeFi-native managers, we have focused on enabling yield strategies, providing early liquidity to various DeFi protocols, and have worked with teams throughout the space on a close basis to grow DeFi liquidity while managing risk.

From our experience deploying capital we bring a unique perspective to risk management as practitioners and power users of DeFi. We have built extensive tooling to provide a deep view into risks in the market including liquidity monitoring, price alerts and a large suite of custom analytics tools. This has allowed us to stay on top of the ever-changing DeFi market, and we bring this expertise to our restaking vaults.

In addition to leveraging our experience for risk management, we bring an extensive rolodex of DeFi teams and have worked with over 500 projects across DeFi for liquidity provision and advisory. This has involved deployment on over 15 chains and dealing with all kinds of conditions within DeFi. We expect to continue this collaboration through the launch of Re7SOL and help expand the restaking ecosystem that is now starting on Solana.

2 Likes
6

This proposal has my support.

While the short-term reallocation of JitoSOL from the DAO treasury presents an opportunity cost, I agree that the long-term benefits of decentralizing the TipRouter outweigh this cost. The additional fees accrued to the DAO through the TipRouter provide a sustainable revenue stream and stronger incentive alignment for ensuring the success of both the Jito protocol, and its restaking platform.

This incentive alignment should allow for better governance oversight of DAO whitelisted vault managers and the fees that fund the DAO itself. It encourages delegates to ensure Jito remains competitive while promoting consistent protocol growth.

Laine raises good questions about the fee structure, and I think it’s important for delegates to periodically assess this to ensure it continues to incentivize new deposits and support protocol growth.

Overall, it’s great to see a core part of Jito’s infrastructure become further decentralized. I trust the DAO to respond appropriately if any of the listed risk or, other are realized.

I will be voting yes on this proposal.

1 Like
7

It should go without saying that Blockworks supports this proposal. The Jito team has done a great job of expanding into new verticals that are complementary to its existing suite of products. We see this proposal as critical to the greater decentralization path of the protocol, as a TipRouter NCN would allow for a decentralized consensus on tip distribution. As we have highlighted in our more recent research pieces, we think that Jito’s revenue potential from this proposal will make the Jito protocol one of the most attractive projects in crypto for the next year. Like @Ian said, this brings better incentives for professional DAO management moving forward.

Given our own models of JIP 8, and recent market activity, we think that there is more potential than meets the eye to this proposal, and so we support this. However, we are curious on the future fee structures to be implemented in this system. From our perspective, Jito has 3 sources of revenue right now: (Re)staking, Liquid staking, and its MEV tips. On a general level, how do others here expect the fee structure to change in relation to the current suite and with possible expansions? Should it be expected that the NCN split will become larger or smaller?

2 Likes
8

As a user and JTO holder,

the percentage to restaking looks small and doubting this attracts users to restake tokens.
I would like to Jito team reconsider those parameters setting.
Restaking users are also taking risks.
Besides I’m concerning only validators are discussing on this forum.
There should be AMA or something to inform to users to involve them.

5 Likes
9

This proposal was edited on 12/1 to reflect a number of major changes detailed here:

While this proposal has completed its mandatory 30-day discussion window, it will not move to vote until there has been sufficient time for the community to discuss these changes.

3 Likes
10

NOTE: This proposal was further edited on 12/17 to include additional information about shared DAO/Developer Council authorities related to the addition of vault managers and node operators, as well as to include information about the maximum number of vault managers and node operators in the initial Tiprouter parameters.

Also includes minor grammatical changes, as well as edits related to DAO treasury assets needed to handle rent obligations.

1 Like