Celer Network White Paper: Generalized State Channels and cStack

Paper title Celer Network: Bring Internet Scale to Every Blockchain
Organization ScaleSphere Foundation Ltd. (“Foundation”), as named in the PDF
Date and status June 15, 2018; the PDF labels itself “Draft, subject to change.”
Authors Mo Dong, Qingkai Liang, Xiaozhou Li, and Junda Liu, as listed in the arXiv bibliographic record. The PDF’s title page names the organization, not these authors.
License No open license is stated in the PDF; this page links to and summarizes the document rather than reproducing it.
Primary source Official Celer Network white paper (PDF)

Educational summary and analysis. CryptoGloss does not claim authorship of the source document or reproduce its text.

Original paper Original paper
Directories Ethereum Whitepapers & Research

The 2018 Celer paper proposes a layered system for moving application activity and value off-chain while relying on an underlying blockchain for security-critical settlement. Its central ideas are generalized state channels, a routing layer for payments across channels, and tools for building off-chain applications. Later products such as the State Guardian Network (SGN), cBridge, and Celer Inter-chain Messaging (Celer IM) developed in separate stages; they should not be read back into the 2018 draft as if they were already operating in its original form.

What the paper proposes

The paper divides the system into four conceptual layers. cChannel provides the off-chain state-transition foundation, using state-channel and sidechain techniques. cRoute addresses how value can move through a network of channels; its proposed Distributed Balanced Routing (DBR) mechanism aims to route payments while balancing channel liquidity. cOS is a framework and runtime for decentralized applications whose state transitions can happen off-chain. cEconomy describes incentive mechanisms intended to support liquidity and service availability. These are the paper’s design proposals, not a guarantee that every design or performance claim was implemented as written.

How a generalized state channel works

A state channel lets participants exchange signed updates without putting every update on-chain. In a typical payment channel, participants establish an on-chain channel and then exchange signed balance states. If they cooperate, many updates can remain off-chain; if they disagree or one stops responding, the channel’s on-chain contracts and dispute process provide a route to settle using an accepted state. This reduces routine blockchain transactions, but it does not remove the need for correct contract logic, valid signatures, and timely dispute handling.

Celer’s design separates payment channels from application channels. The current CelerCore channel documentation describes CelerPay as the payment-channel layer and CelerApp as a way to express application logic and outcomes. CelerNodes run the protocols. CelerApps can be connected to conditional payments, so an application result can determine whether a payment completes. The documentation also describes virtual application channels that can avoid an on-chain initialization when participants cooperate; payment channels and disputes still have their own on-chain requirements.

Routing across channels

A channel connects a limited set of participants, so a network needs a way to route a payment through intermediaries. The paper’s cRoute section proposes DBR to find routes while accounting for available balances across channels. Multi-hop routes introduce practical constraints: a path must have sufficient liquidity, intermediate nodes must be online and able to forward, and conditional payments must resolve safely. The paper’s simulated or theoretical performance claims should be understood in the context of its stated model, rather than treated as measured current network performance.

How later Celer services differ

State Guardian Network (SGN): Celer’s later documentation describes SGN as a specialized proof-of-stake chain. In state-channel contexts, it can help store channel-state information and respond to settlement activity when needed. The SGN also took roles in later services. Its current responsibilities should be described from the relevant product documentation, not inferred solely from the 2018 paper.

cBridge: Celer introduced cBridge in 2021 as a cross-chain value-transfer service. Its documentation distinguishes xAsset, a canonical-mapping model in which assets are locked and corresponding representations are issued or redeemed, from xLiquidity, a pool-based model in which users deposit into a source-chain pool and receive assets from a destination-chain pool. The SGN coordinates parts of the transfer flow. Fees, liquidity limits, pool imbalance, contract behavior, and the specific route matter; it is inaccurate to describe every cBridge transfer as one universal “liquidity relay” or “lock-and-mint” process.

Celer IM: The Inter-chain Messaging framework lets applications send messages between chains through MessageBus contracts and SGN processing. Celer’s developer materials describe validators reaching consensus on source-chain messages before an executor submits them for destination-chain execution. Applications can add their own guardian safeguards. This architecture has different trust and failure considerations from a two-party state channel, so the security model depends on the messaging route and the application’s configuration.

Tradeoffs and what to verify

Off-chain systems can reduce the frequency of on-chain transactions, but they shift some work to participants, routing nodes, contracts, and monitoring services. Users may need channel liquidity, reliable connectivity, and a workable dispute path. Bridge and messaging users also need to understand which mechanism applies to a specific transfer and what happens if a validator, relayer, contract, or liquidity pool fails. The 2018 paper is useful for understanding Celer’s early architecture; it is not a current product manual, an audit, or evidence that every proposed feature shipped unchanged.

Related Terms

See Also

Sources