TIP-44: Launching Treasure L2 on ZKsync

Proposals
1

Recovered from https://forum.treasure.lol/t/tip-44-launching-treasure-l2-on-zksync/80 Originally posted 2024-09-12.

image1000×562 449 KB

Summary

This note has been published on behalf of the Treasure core contributor team.

The proposal from Matter Labs and the accompanying proposal issued jointly by both ZKsync Foundation and the Treasure core contributors revisits the ratified proposal TIP-36: Ecosystem Selection for the Treasure Chain (approved by TreasureDAO on February 21, 2024) to launch the Treasure L2 as the leading gaming chain on ZKsync’s Elastic Chain, instead of on Arbitrum Orbit.

This TIP-44 combines the two proposals from Matter Labs and ZKsync Foundation, as well as outlines the final parameters for the launch of the Treasure L2 across technical specifications and timeline.

Background

This was an unplanned path for Treasure’s core contributors but unexpected turns often lead to the best outcomes. Since TIP-36, we have been full steam ahead and committed to launching the Treasure L2 on Arbitrum Orbit, working closely with our tech and infrastructure partners, actively participating in Arbitrum DAO governance as one of its largest delegates, and continuing to be a strong and avid supporter for Arbitrum gaming through our efforts to bring games to Treasure and conceive the Arbitrum Gaming Catalyst Program (GCP).

As we approached this critical point ahead of the launch of Treasure mainnet before this “point of no return” and other events that have ensued, we decided to take a step back and do a “cold shower” reflection on the future of Treasure, asking ourselves the same question we had last year: “Is Arbitrum the best home for Treasure?”

Complacency has never been part of Treasure’s culture. We always want to go into every decision armed with as much confidence as possible. Therefore, we decided to conduct an assessment on Arbitrum across all areas, specifically governance. We reflected on the past few months following TIP-36, the state of GCP, and the amount of time and energy devoted to the program compared to its results so far. Treasure has been a pillar of Arbitrum DAO governance to-date and led the way in establishing a sub-DAO to enable greater community governance. After these reflections, we decided to revisit earlier conversations with other ecosystems to better understand their vision, ethos, values, technology and product roadmap, and level of fit with Treasure as a consumer-focused ecosystem.

It is through this exploration that we have come to the conclusion that ZKsync is the best home for the Treasure ecosystem.

Over the past four weeks, we have worked closely with the Matter Labs and ZKsync Foundation teams to explore a partnership and prepare this proposal to present the opportunity to the DAO. As Treasure’s core contributors, we are always operating with the best interests of the Treasure ecosystem at heart – this spans the future of the network, our game partners and builders, and its end users. We believe that the best path forward for our community is to launch the Treasure L2 on ZKsync as part of the Elastic Chain.

Below we outline the reasoning behind our conviction.

Rationale

2

Scaling Ethereum will involve many tech stacks with unique advantages and disadvantages. Certain projects will fit better on different chains. ZKsync’s technical roadmap fits our vision for where Treasure can go.

Treasure is the decentralized game console. We are building a new entertainment system that marries fun and novel gaming experiences, a passionate and tight-knit community, and a powerful builder-player flywheel that breaks down traditional walled gardens to align interests across the network. All of this is connected by MAGIC.

To achieve this goal, Treasure needs to be built atop a future proof technology solution within an environment tailor-made to meet the infinitely expanding demands of mass adoption and that is built with the consumer experience at its very core. ZKsync checks off all of these boxes and more to unlock new paradigms for us to push the boundaries of what’s possible, together.

ZK is the clear endgame for rollups as said by Vitalik Buterin himself. Put simply, ZK Rollups differ from Optimistic Rollups (like Arbitrum) in that they are based on proven correctness versus assumptions, allowing for instant finality while also enabling eventual higher throughput and lower costs as transactions are compressed into succinct proofs (vs. raw data). Eventually, this can allow blockchains to scale to meet near “infinite” demand. DeFi and gaming have very different starting principles. DeFi emphasizes openness and transparency while web3 games are highly contingent on some degree of opaqueness. While building on current rollups, we have often encountered difficulties in creating ideal gaming experiences in open information environments. Matter Labs is not only focused on the kinds of problems we are trying to solve in a different context. They are ingrained with bleeding edge efforts to accelerate zero-knowledge technology that will prove crucial to making better web3 games.

To learn more, visit the ZKsync website and see this interview with Alex Gluchowski, the Co-Founder and CEO of Matter Labs. image902×924 114 KB

3

ZKsync’s ZK Stack is the only stack that supports Native Account Abstraction which is massively important for consumer use cases. This unlocks a new frontier for more seamless blockchain abstraction without bothering consumers with technicalities – a huge enabler that will allow builders and game developers within Treasure to offer better gaming and consumer experiences suitable for the masses. In the ZKsync world, all accounts (including Externally Owned Accounts (EOAs)) are considered to be smart contract accounts that enable features like gas sponsorship and the ability to pay gas fees with ERC-20 tokens. With Treasure Connect and our Global Treasure Accounts system, this will eliminate onboarding friction and remove onchain interaction barriers across all networks. We no longer need to imagine a world where you can send $MAGIC to a recipient using gamer@treasure.chain, or make a contract call using magicswap@treasure.chain – thanks to future initiatives and partnerships, this will be a reality and available on ZK Chains in the near future.

Under the hood, Native AA is built around a super-set of the highly extensible ERC-7579 modular [minimal] account standard, forming what is generally known as ZK Accounts. As a minimal standard, ERC-7579 is highly gas efficient and affords account builders limitless extensibility via modules. Various community/public modules have emerged and Treasure will innovate further with novel integrations across our games, marketplace, AMM and more.

This is coupled with a diverse set of Paymaster strategies (e.g. general, token-gated, approval-based) that can be leveraged to craft custom gas sponsorship policies, orchestrated on a per-transaction and per-wallet basis via the Treasure Gas Sponsorship Service. Strategies can be applied to both Treasure Smart Accounts and vanilla EOAs, to completely streamline user onboarding - a significant advantage.

4

ZKsync’s Elastic Chain is an infinitely extensible network of ZK Chains seamlessly interconnected with hyperbridges - the native and trustless communication protocol of the ZK Stack. It enables any chain to natively exchange messages and transfer value with any other chain, thus tapping into the users and liquidity of the entire ZK chain ecosystem. Unlike classic, centralized bridges, hyperbridges rely solely on cryptography for security.

This is the ultimate expression of interoperability and to date the clearest path we have seen towards building the decentralized gaming console that spans all Treasure Chains. To users, the Elastic Chain can be seen as a single blockchain encompassing the entire network to offer true interoperability between ZK Chains. Chains can interact and transact with each other efficiently enabling games and protocols deployed on one chain to be interacted with across any other chain, effectively eliminating the need for traditional bridging.

This is a powerful unlock for Treasure games to not only leverage shared universal liquidity and assets, but any protocol deployed to the Elastic Chain. We expect this to be a key differentiator for a superior developer and user experience.

5

The ZKsync ecosystem today features other consumer-focused chains and networks such as Abstract (the same folks behind Pudgy Penguins) and Lens (decentralized social network) among others who, like us, are working to create and support consumer experiences that can bring billions onchain.

The capability for real interoperability and unified liquidity between chains within the Elastic Chain coupled with other consumer networks being built within ZKsync unlocks real network effects that can better support the Treasure ecosystem.

While it is irrefutable that Arbitrum currently has the most bridged ETH and TVL in contrast to other chain ecosystems, much of this liquidity rests within DeFi that, generally, is comprised of more passive users optimizing for the best yield. Since Treasure’s inception, we have converted more non-DeFi users to Arbitrum than the other way around.

We believe in more positive sum games versus zero sum, especially as we transcend building consumer products and experiences that go beyond crypto users to onboard the masses to Treasure. This same modus operandi is shared by other members of the Elastic Chain. This spirit as well as the ability to better cooperate will support the success of the overall system.

6

Earlier this year, EIP-4844 brought Data Availability optimization for chains (Arbitrum, Optimism, etc.) and associated cost reduction due to the ability to post larger rollup blobs. This however is a temporary solution as, at some point, blob space will again get saturated, resulting in a steep and rapid increase to transaction costs across all optimistic rollups. This is guaranteed with web3 on the cusp of achieving mainstream consumer adoption.

ZK Chains will be impervious to this to allow for greater transaction price sustainability and stability. Fees will trend downwards only, as state diffs’ compression technology advances.

In addition to DA, execution capacity is another significant bottleneck which most rollups are not addressing. ZKsync is the most mature with regards to actual throughput capacity and how parallel processing, batching, prover and ZK cryptography can be leveraged for optimization.

When it comes to interoperability, ZK Chain Hyperbridges bring additional cost savings that are less tangible and not often talked about. With classic bridges, costs encompass intermediaries’ fees on top of transaction fees. All fees other than transaction fees are eliminated on ZK Chains; a small reduction on a per-transaction basis but massive on a cumulative scale.

7

Per the proposal penned by Alex Gluchowski, Matter Labs will look to partner with Treasure to operate our L2 blockchain powered by ZKsync technology. As the core contributors behind ZKsync and the ZK Stack, we will be able to count on Matter Labs as our technology partner to provide best-in-class support to launch, operate, and maintain the Treasure L2. This will allow us to benefit from future upgrades, customizations, and DevOps support for the benefit of the Treasure ecosystem.

ZKsync Foundation’s token swap has been structured to allow for shared, long term alignment between our respective ecosystems. As a fair launch DAO, Treasure has no form of traditional equity. Maintaining alignment at the token level is crucial for us to ensure that the ZKsync ecosystem prioritizes Treasure and that we can support each other’s success. This same alignment will stream down to our game partners and all other builders on Treasure as well, allowing for a rising tide that will lift all boats.

8

After speaking with the ZKsync Foundation, the ZKsync Association, and other delegates in the ZK Nation, we believe they are set up for success with strong alignment around a core mission and a focus on local consensus versus global consensus (thus minimizing voting while encouraging max governance participation). Importantly, they are striving for efficient and meritocratic outcomes driven by mechanisms over committees. This will mean that builders can remain focused on building and scaling their projects over needing to participate in governance to gain access to funding and support. With Treasure’s governance infrastructure and expertise through the ARC, there is an opportunity for us to be early to apply our experience to the ZK Nation and contribute to standing up a governance system and environment that is more builder-centric.

9

We have been disappointed with the pace of the Arbitrum Gaming Catalyst Program (GCP) to begin catalyzing and supporting gaming on Arbitrum. Where we are at today was a concern that has been flagged, urging Arbitrum and the DAO to move with a greater sense of urgency (see posts from @AvgJoesCrypto on Jan 30 and @karelvuong on Feb 27 / Mar 14 / Aug 7).

image1000×367 60 KB

As one of the original contributors to the GCP’s conception, we believe that, while well-intentioned, the DAO has contributed to the GCP being too slow-moving. On August 7, Karel Vuong (Co-Founder of Treasure) resigned from the GCP Working Group to eliminate some confusion on his role as stated here. There are great members of the GCP Working Group and Council who we believe will be able to do some amazing work to support gaming as the program continues to progress.

Speaking with our game partners as well as new games who we have been in talks with to join Treasure, many share similar feelings.

Impact to Users and Builders

With Treasure planned to have been launched as an L2, our goal has always been to build a decentralized gaming ecosystem with a strong community, leading tech stack, and purposeful culture that sits atop Ethereum. Whether the Treasure L2 was built and powered by Arbitrum Orbit, ZK Stack, OP Stack, or another rollup stack, we have been focused on making this transition to Treasure as seamless as possible for both users and builders.

For users:

  • Asset Bridging - No difference in either case. Users would have been required to bridge assets over from Arbitrum One to Treasure regardless of the chain being launched on Arbitrum Orbit or ZK Stack. The Treasure core team has been focused on ensuring that the bridging and migration process is as seamless as possible to move assets over from Arbitrum One ($MAGIC, ERC20s, and ERC721s/ERC1155s).

  • Transaction Costs - Initially higher on ZK Stack due to offchain fees relating to ZK proving. This fall, we expect this to be optimized to <$0.01 with continual long-term cost reductions driven by advancements in prover technology.

  • Apps/Protocols:
    Magicswap - No difference in either case. Pools will be relaunched on the Treasure L2 with migration incentive schemes implemented to drive existing liquidity to Treasure from Arbitrum One as noted in TIP-43: Chain Migration Incentives.
    Treasure Market - No difference in either case. The Treasure Market (NFT marketplace) would be redeployed on the Treasure L2 to support NFT trading for all migrated collections.
    Treasure Games - No difference in either case. Treasure’s games have signalled their excitement and commitment to migrating to the Treasure L2 over time regardless of the underlying tech stack.

For builders:

  • Users/Distribution - In either case, Treasure will be migrating its existing user base from Arbitrum One to the Treasure L2 with a seamless bridging experience. With the Elastic Chain’s interoperability, Native Account Abstraction paired with our Global Treasure Account system, other consumer networks that will help cross-pollinate, ZKsync will supercharge our existing plans to onboard the masses into Treasure.

  • Liquidity - While Arbitrum is currently ahead of the ZKsync ecosystem on metrics of bridged ETH and TVL, much of this liquidity is captured within DeFi applications and generally immobile, not benefiting consumer applications and ecosystems. With ZKsync’s Elastic Chain offering a shared liquidity layer, more seamless interoperability capabilities across ZK Chains, and a major focus on growing consumer networks (such as Treasure), we expect ZKsync to accelerate in overall adoption in the near future.

  • EVM Compatibility - Treasure on ZK Stack is fully EVM Compatible and most existing smart contracts can be deployed without any modifications. There are some differences that exist which are summarized here. ZKsync is actively working towards achieving Type-1 EVM Equivalence (currently Type-4).

  • Gas Sponsorship - Initially higher on ZK Stack due to offchain fees relating to ZK proving with continual cost reductions driven by improvements in prover technology (starting as early as this fall).

  • Treasure Platform and Tooling - No difference in either case. We are working to ensure Treasure Connect, TDK, Magicswap, Treasure Market, and other Treasure infrastructure is supported on the Treasure L2.

  • Vendor and Ecosystem Services - No difference in either case. All previously planned Treasure L2 vendors and protocols are supported with additional options now available who are focused purely on harnessing ZK tech to enable new frontiers for development.

  • Account Abstraction - As Treasure Connect and Treasure Accounts will serve as the defacto account abstraction solution for apps and games building atop the Treasure L2, ZK Stack’s Native Account Abstraction offers additional unlocks that will enable Treasure and developers to offer superior consumer onboarding capabilities.

  • Stylus - Launched on September 3, 2024, Arbitrum Stylus is exclusive to Arbitrum Orbit and will not be available for use on ZK Stack. Stylus offers EVM+ Compatibility, allowing developers to develop smart contracts not only in Solidity and Vyper, but also Rust, C, C++, and more.

  • Infinity Chains (L3s) - ZK Stack will offer additional benefits to Infinity Chains arising from interoperability unlocks via ZKsync’s Elastic Chain. This will allow for Treasure to achieve its vision of users being able to seamlessly glide between games (and game-specific chains), as well as other ZK Chains within the Elastic Chain.

10

Notes:

  • Other ecosystems have been omitted for the purposes of this comparison between Arbitrum and ZKsync.

  • Additional metrics can be found at growthepie.xyz, DefiLlama, and L2Beat.

Criteria
Arbitrum (Orbit)
ZKsync (ZK Stack)
Rollup Differences
Optimistic Rollup
ZK Rollup
Stack Maturity
Mature. Dozens of Arbitrum Orbit chains in mainnet.
Rapidly maturing. Growing number of ZK Stack chains in mainnet.
EVM Compatibility
Fully compatible with additional Stylus support for EVM+ Compatibility.
Type 4 (High-Level Language Equivalence). Type 1 (Full EVM Equivalence) is planned for early 2025. See differences here. Most existing smart contracts on Ethereum or Arbitrum can be deployed without any modifications.
Development Languages
Solidity, Vyper. Rust, C and C++ via Stylus.
Solidity and Vyper.
Transaction Costs
$0.004 in last 24 hours - pricing will remain fairly static over time with marginal improvements due to technological-economical bounds. See fee mechanism.
$0.016 in last 24 hours - current higher floor due to offchain fees relating to ZK proving; <$0.01 this fall with continual long-term cost reductions driven by advancements in prover technology. See fee mechanism.
Custom Gas Token
Supported.
Supported.
Interoperability
Chain Clusters (under going research). Interop between Orbit stacks will see finality improvements by balancing trust assumptions and leveraging secondary consensus protocols.
Interoperability through Shared Bridge in the Elastic Chain. Consensus by cryptographic verifiable computation - ZK eliminates trust assumptions enabling novel implementations.
Data Availability (DA) Customization
Several options exist (Rollup, DAC, and AltDA such as EigenDA).
Several options exist (Rollup, Validium, Volition, zkPorter and AltDA such as EigenDA).
L3 Support (for Treasure’s Infinity Chains)
Yes, via Arbitrum Orbit.
Yes, via ZK Stacks (extension to Elastic Chain).
Account Abstraction
App-layer. Dependency on deployed AA and Paymaster protocols (e.g. ERC-4337).
Native. Unparalleled UX and flexibility. Smart accounts & paymasters are first-class citizens; EOAs can behave as smart accounts and have paymaster support.
TPS
896 (highest recorded); 40,000 (theoretical max)
200-500 (single stack); eventual 6-figure upper bounds across the Treasure network.
Development Frameworks
Hardhat, Foundry, Truffle, thirdweb, Brownie
Hardhat, Foundry (fork)
Revenue Share
Arbitrum Orbit Expansion Program (10% of sequencer fees)
ZK Protocol Fees - not yet launched (expected to be 10% of sequencer fees)
Ecosystem
Wide variety of vendors and protocols
All previously planned Treasure Chain vendors and protocols supported with additional options focused purely on harnessing ZK tech

Proposal

This proposal merges the two posts submitted from Matter Labs and ZKsync Foundation, as well as outlines the final technical specifications for the Treasure L2 to:

  • (A) Launch the Treasure chain as an L2 leveraging ZK Stack, instead of Arbitrum Orbit.
    Ruby Sunset: Ruby, Treasure’s previous testnet built on Orbit will remain in operation for a limited time as protocols and partners redeploy to Topaz (first Treasure testnet leveraging ZK Stack).
    Data Availability (DA): Topaz will use a Validium Data Availability strategy. For mainnet, we aim to migrate to EigenDA when custom quorum and ZKsync support comes to EigenLayer.
    Gas token: Both Topaz and mainnet will have $MAGIC as the native gas token.
    Faucet, Bridge & Explorer: Topaz and mainnet deployments will each include faucet, asset bridge(s) and block explorer.
    Deployed Protocols: All previously planned partner protocols and developer tools will be deployed and supported shortly after Topaz and mainnet undergoes genesis. This includes Thirdweb, Goldsky, LayerZero, Palmera & Safe, Chainlink CCIP, EAS and more.
    Permissionless: Treasure will be launched as a permissionless network, allowing any developer to deploy smart contracts at-will (in contrast to initially being via ACL as indicated in the litepaper).
    Native Accounts: Treasure’s smart accounts will migrate to leveraging ZKsync’s Native Account Abstraction solution as part of the transition.

  • (B) Target the launch of Treasure chain mainnet within the next 1.5-2 months (latest by mid-November).

  • (C) Execute the token swap between TreasureDAO and ZKsync Foundation on the following parameters:
    Swap Ratio: 3.19:1 of $ZK to $MAGIC
    5,264,879 MAGIC tokens moved from the TreasureDAO Ecosystem Fund (arb1:0x482729215AAF99B3199E41125865821ed5A4978a) (556,739 from Y2, 2,226,958 from Y3, 2,226,958 from Y4, and 254,224 from Y5) in exchange for 16,800,000 ZK tokens
    Milestones: Token swap is conditional on the successful launch of Treasure mainnet on ZK Stack.
    Lockup and Vesting: Both TreasureDAO and the ZKsync Foundation will adhere to a symmetrical 12-month token lockup, followed by a 24-month gradual unlocking schedule.

  • (D) Wind down the Treasure ARC (Arbitrum Council)
    Per TIP-38, the current term for Treasure ARC’s Council Members spans through to October 30, 2024. If TreasureDAO opts to move to ZKsync, the mandate of the ARC will need to be revisited. A governance proposal will be submitted by the ARC on this in the near future.

The following options are planned to be presented to the DAO during the Snapshot voting period:

  1. Approve (Migrate to ZKsync)

  2. Reject

  3. Abstain

Next Steps

The temperature check and community discussion period begins now and will end September 17, 2024 at 20:30 UTC (UNIX timestamp: 1726605000). If quorum is reached, a subsequent Snapshot vote will be posted shortly after.