Skip to main content

FAQ

How is this different from shared sequencing?

A shared sequencer is a centralized entity that controls transaction ordering across all rollups, creating bottlenecks and forcing rollups to give up sovereignty. We only coordinate cross-chain transaction inclusion - each rollup maintains its own sequencer and full control over transaction ordering.

What happens if the shared publisher goes down?

Our "based" architecture provides multiple layers of resilience:

  • Automatic Failover: Publisher role rotates with Ethereum's block proposer schedule - if current publisher fails, next L1 proposer automatically becomes publisher
  • Ethereum Liveness Dependency: Publisher availability tied to Ethereum itself (99.99% historical uptime)
  • Graceful Degradation: Rollups continue independent operation and can resume cross-rollup and settlement functionality when publisher comes back online
  • No Data Loss: Failed cross-rollup transactions abort atomically - no partial states or stuck funds
  • Permissionless Recovery: Any Ethereum validator can step in as publisher through L1 proposer rotation

How do you maintain atomicity across rollups?

We use Two-Phase Commit (2PC), the gold standard for atomic transactions in distributed databases:

  • Phase 1 (Prepare): Publisher asks all rollups "Can you execute your part?" Each rollup simulates execution and votes commit/abort
  • Phase 2 (Commit/Abort): Only if ALL rollups vote commit does the publisher send execute command. If anyone votes abort, all rollups abort atomically
  • Proven at Scale: Same protocol used by Oracle, PostgreSQL, and other major databases for decades
  • Failure Handling: Network partitions and rollup failures result in atomic abort - no partial execution states

What's required for rollup integration?

Integration is deliberately meant to be minimal and non-invasive:

  • Mailbox Smart Contracts: Simple contracts for cross-rollup message reading/writing (similar to existing bridge interfaces)
  • 2PC Participation: Rollup sequencer participates in prepare/commit voting protocol with shared publisher
  • Message Interface: Implementation of standardized message-passing protocol
  • Minimal Changes: Sovereignty over sequencing logic, consensus mechanisms, or governance structures
  • EVM Compatible: Works with EVM, other architectures are TBD
  • Existing Infrastructure: Leverages current rollup proof systems, DA layers, and settlement contracts

How does this affect existing rollup economics?

Rollups keep 100% of their existing fee revenue and sequencer economics. Cross-rollup functionality can actually increase transaction volume and fees by enabling new types of applications and user flows.

Is this just another bridge solution?

Fundamentally different from bridges in both architecture and capabilities:

Bridges: Point-to-point asset transfers with multi-step processes and failure points

Composable: Native cross-rollup execution infrastructure enabling complex atomic and synchronous workflows

For example:

Bridge Example: Move USDC from Rollup A to Rollup B (separate transactions, bridging delays, trust assumptions)

Our Example: Atomic swap of ETH on Rollup A for USDC on Rollup B in single transaction with immediate finality

Bridge Risk: Protocol vulnerabilities, liquidity requirements, and multi-step failure modes

Our Approach: Built into rollup infrastructure with database-proven atomic guarantees and no additional protocols

What about performance and latency?

Designed for superior performance through architectural advantages:

  • Horizontal Scaling: Enables parallel processing across rollups rather than sequential bridge operations
  • Single L1 Settlement: All cross-rollup activities are settled within a single Ethereum block
  • Immediate Finality: Once published to L1, there are no additional confirmations needed, providing immediate finality
  • Minimal Overhead: The 2PC coordination system introduces negligible latency compared to multiple bridge hops
  • Parallel Execution: Independent rollup sequencers can process transactions concurrently