Glossary · Scaling
Published 22 May 2025 · Updated 9 Nov 2026 · 16 min read
Layer 2 (L2) is a separate network that processes transactions away from a base blockchain while relying on it to protect users' funds. Designs publish data or signed commitments so Ethereum or Bitcoin can check results or settle disputes. Each design makes different demands on users and operators.
A blockchain has limited room because every node must check its work. Layer 2s move execution elsewhere while keeping a claim on that chain's security. The base network, such as the Ethereum blockchain, anchors settlement rather than doing every calculation. The definitions below show why different sources draw the boundary differently. Some describe blockchains specifically, while others also include payment channels.
| Source | Date | Definition, in its own words | What it adds |
|---|---|---|---|
| ethereum.org, What is layer 2? | DatePage updated 14 October 2026 | Definition, in its own words“A layer 2 is a separate blockchain that extends Ethereum and inherits the security guarantees of Ethereum.” | What it addsSecurity inherited from the base chain |
| L2BEAT, now ETHBEAT, glossary | DateRead 9 November 2026 | Definition, in its own words“A blockchain that inherits consensus and data availability from another blockchain called L1” | What it addsIts definition of a rollup, the strong kind of layer 2 |
| Vitalik Buterin, Different types of layer 2s | Date31 October 2023 | Definition, in its own words“what level of guarantee do you have that you will be able to take the asset back to the L1?” | What it addsThe test is getting your assets back |
| Joseph Poon and Thaddeus Dryja, Lightning white paper | Date14 January 2016 | Definition, in its own words“transactions are sent over a network of micropayment channels (a.k.a. payment channels or transaction channels) whose transfer of value occurs off-blockchain.” | What it addsPayment channels on Bitcoin |
| BIS Bulletin No 126, appendix glossary | Date6 July 2026 | Definition, in its own words“solutions built on the layer 1 network, where most transactions occur off-chain and are only sporadically or periodically reported back to the underlying layer 1 network.” | What it addsA BIS staff view that sets sidechains apart |
| ITU-T X.200 (identical to ISO/IEC 7498-1) | Date1994 | Definition, in its own words“The Data Link Layer detects and possibly corrects errors which may occur in the Physical Layer.” | What it addsThe networking meaning: layer 2 of the OSI model |
Test where data lives, how results are checked, and whether users can leave without operator permission. With blockchain rollups, data is published on L1, which checks or challenges state. L2BEAT, renamed ETHBEAT on 4 November 2026, calls rollups strong L2s and validiums and optimiums light L2s. The latter keep data elsewhere. Its 19 June 2025 rule requires working proof systems and a data availability bridge. Projects failing either requirement go under "Other".
Sidechain networks use their own consensus and connect to another blockchain through a two-way bridge. Their validators or federation, rather than the original chain's proof rules, can approve an invalid state. Bitcoin Layers sorts systems on the Bitcoin network into bitcoin native systems, sidesystems and alternative chains. Its layer 2 definition requires a unilateral exit through a Bitcoin transaction. Lightning meets it because each participant can close a channel independently. A connection between networks does not itself transfer security guarantees.
In networking, layer 2 means the data link layer of the OSI model, not a blockchain scaling system. ITU-T X.200, identical to ISO/IEC 7498-1, describes this separate meaning. It groups bits into frames and "detects and possibly corrects" errors from the physical layer. Devices use a 48-bit MAC address on a local network. Ethernet follows IEEE 802.3, and a layer 2 switch forwards frames using the destination MAC address.
A deposit, a confirmation and a withdrawal are different events. Follow a rollup transaction through six steps to see which party acts at each point. Most rollups run one sequencer in November 2026. Separate its acknowledgment from the checks that make the result enforceable.
Users deposit through the canonical bridge, whose base chain contract checks proofs. Third-party bridges add trust in their own signers rather than just the rollup's verification rules.
A sequencer orders transactions. Forced inclusion offers a route through L1 if it ignores them. In November 2026, delays reach one day on Arbitrum One and seven days on Scroll.
The rollup executes transactions away from Ethereum. A fast confirmation is the sequencer's promise, not final settlement. The base chain has not yet accepted that result.
Dencun introduced EIP-4844 blobs for transaction data in March 2024. Each holds 131,072 bytes for about 18 days. PeerDAS, introduced in December 2025, lets nodes check availability by sampling.
An optimistic rollup design accepts state unless a fraud proof succeeds during its challenge period. Arbitrum One allows 6 days 8 hours in November 2026. A ZK rollup requires a validity proof with each state update.
Withdrawals follow the challenge period or verified proof. An escape hatch enables an independent exit through L1. The exit window is the time users have to leave before an unwanted upgrade takes effect. Emergency upgrade paths can remove it.
Other scaling designs take different routes. A payment channel normally needs only opening and closing transactions on the base chain. The Plasma design posts commitments rather than full transaction data. If an operator withholds that data, users must exit because they cannot assemble the evidence needed to challenge it. Publishing transaction data on Ethereum addresses that specific problem by making independent checks possible.
The figure applies three security tests to eight systems on Ethereum and Bitcoin. Select a system, then compare normal operation with an operator outage. The 9 November 2026 snapshot checks data location, verification and the user's independent route out of an L2.
Arbitrum One posts data on Ethereum with fraud proofs. This Stage 1 optimistic rollup allows self-sequencing after up to one day.
Scroll posts data on Ethereum with validity proofs. This Stage 0 rollup allows forced transactions after up to seven days, but no pre-upgrade exit window.
Lens uses validity proofs but posts data on Avail. Withdrawals freeze if its proposer fails, despite the proof system.
Celo uses fraud proofs with data on EigenDA. Withdrawals freeze if its proposer fails, and funds can be lost if that data becomes unavailable.
Polygon PoS validators sign state without a fully functional proof system. ETHBEAT files this sidechain under "Other", and ethereum.org lists it as a sidechain.
Lightning peers keep channel data themselves. Bitcoin enforces the latest signed state if someone watches, and either peer can close alone.
Rootstock secures its own blocks through merged mining. BTC exits depend on its 5-of-9 PowPeg federation, not a user acting alone.
Citrea calls itself a rollup with data on Bitcoin. Its bridge needs one honest participant per key role, and users cannot withdraw alone.
This figure summarizes each system's published design and ETHBEAT's assessment of 9 November 2026.
The Ethereum examples separate three dependencies in the 9 November 2026 assessment. Arbitrum One and Scroll combine published data with working proofs. Lens and Celo keep data elsewhere, creating validium data risks or equivalent optimium risks. A data availability committee or external provider adds trust beyond verification. Withheld data can leave funds frozen or lost. Polygon PoS instead depends on validator approval, so ETHBEAT places this sidechain under "Other" rather than treating it as a rollup.
Exit permissions matter especially for the Bitcoin cases checked on 9 November 2026. Lightning keeps state with peers yet supports unilateral closure. Rootstock is a merge-mined sidechain with a 5-of-9 federation controlling BTC exits. Citrea describes Bitcoin as its data and settlement layer. Its Clementine bridge uses BitVM2 and needs one honest participant in each key role, but users cannot exit alone. Independent risk ratings cover Bitcoin systems far less than ETHBEAT covers Ethereum ones, so the Bitcoin rows rest mostly on each project's own documentation.
A blockchain state channel lets two participants update balances without recording every payment on a blockchain. The Lightning Network applies this pattern to Bitcoin. Signed states give either peer a way to enforce the balance independently.
One on-chain funding transaction locks the channel's funds in a 2-of-2 multisig output, which needs both signatures to spend.
The peers sign an updated commitment transaction and revoke the previous one. This payment adds no transaction to Bitcoin.
An HTLC carries the payment through a middle node. The recipient supplies a preimage, and the forwarding node earns its routing fee.
A cooperative close pays both participants through one on-chain transaction. Neither side faces the unilateral close's extra delay.
A unilateral close broadcasts the latest commitment transaction. The closer waits to_self_delay blocks before spending its own output.
Publishing a revoked state allows a penalty transaction taking all the cheater's outputs, if the peer or a watchtower acts in time.
Splicing changes a channel's size without closing it. It was merged into the specification in March 2026, and LND had no splicing release in November 2026.
This walkthrough follows the Lightning specification and each client's default settings checked on 9 November 2026.
With defaults checked on 9 November 2026, each hop charges 1,000 millisatoshi plus a rate per million. Rates are 1 for LND, 10 for Core Lightning and 200 for Eclair's public channels. The calculated cost for a 100,000 sat payment is 1.1, 2 or 21 sat per hop. Force-close delays are 144 blocks, about a day, for Core Lightning, 720 for Eclair and 144 to 2,016 for LND, depending on channel size. Splicing resizes channels without closing them. It was merged into the specification in March 2026.
Capacity depends on what each transaction requires, not just a headline rate. Compare a data-only model with Ethereum's execution ceiling of about 238 plain transfers per second on 9 November 2026. The calculation is not a throughput measurement.
At Dencun's March 2024 target of three blobs, the model carries 2,731 12-byte transfers per second or 182 180-byte ERC-20-sized transfers.
Pectra doubled the target to six blobs in May 2025, doubling modeled data capacity at the same transaction size.
Since January 2026, 14 blobs per block model 3.6 to 53.5 times L1 execution, using 12 to 180 bytes per transaction.
The November 2026 maximum allows 21 blobs per block for high demand. It is a ceiling, not the target.
These page team calculations use Ethereum's published parameters, Vitalik Buterin's byte estimates and an OP Stack fee floor, not measured throughput.
Blob targets rose from 3 in March 2024 to 14 in January 2026, but layer 2 execution and proving remain separate limits. Blobs carry data, not computation. Blobscan's hourly counts give 7.00 blobs per slot over the 24 hours to 20:00 UTC on 9 November 2026, half the target. ETHBEAT's past-day reading that date shows about 1.04 thousand user operations per second on rollups, about 443 for "Others", and 21.67 for Ethereum. Operation counts do not represent equal work.
The table separates mechanisms that are often grouped together in scaling discussions. Read across a row to connect data storage, verification and exit rules. The sidechain row is a contrast, not another member of the layer 2 family.
| Family | Where data goes | How state is checked | How users exit | Typical withdrawal | Who can stop or censor | Examples |
|---|---|---|---|---|---|---|
| Optimistic rollup | Where data goesOn the base chain, in blobs | How state is checkedFraud proofs within a challenge window | How users exitWithdraw after the window, or force a transaction through L1 | Typical withdrawalAbout a week (6 days 8 hours on Arbitrum One) | Who can stop or censorThe sequencer can delay, users can go through L1 | ExamplesArbitrum One |
| ZK rollup | Where data goesOn the base chain, in blobs | How state is checkedA validity proof with every state update | How users exitWithdraw once the proof is verified | Typical withdrawalHours to a day | Who can stop or censorThe sequencer can delay, forced inclusion varies | ExamplesScroll, ZKsync Era, Starknet |
| Validium | Where data goesOff the base chain, with a committee or a data layer | How state is checkedValidity proofs | How users exitNeeds the data that is kept elsewhere | Typical withdrawalLike ZK rollups when data is available | Who can stop or censorWhoever holds the data can freeze funds | ExamplesLens (data on Avail) |
| Optimium | Where data goesOff the base chain, with a committee or a data layer | How state is checkedFraud proofs | How users exitNeeds the data that is kept elsewhere | Typical withdrawalLike optimistic rollups when data is available | Who can stop or censorWhoever holds the data can freeze funds | ExamplesCelo (data on EigenDA) |
| State channel | Where data goesWith the channel's peers | How state is checkedSigned states, penalties for old ones | How users exitEither peer closes the channel alone | Typical withdrawalA cooperative close once it confirms, or a delay of about a day or more | Who can stop or censorA peer can go offline, someone must watch | ExamplesLightning Network |
| Plasma | Where data goesOnly commitments on the base chain | How state is checkedFraud proofs that need the operator's data | How users exitUsers exit if the operator withholds data | Typical withdrawalA challenge period, mass exits if data is withheld | Who can stop or censorThe operator can withhold data | ExamplesPlasma designs of 2017 and later |
| Sidechain (for contrast) | Where data goesOn its own chain | How state is checkedIts own consensus, no proof checked by the base chain | How users exitThrough a bridge its validators or federation run | Typical withdrawalSet by the bridge | Who can stop or censorThe validators or federation | ExamplesPolygon PoS, Rootstock |
Data availability determines which parties users must trust to reconstruct balances. A proof system determines whether an operator can settle an invalid state. Understanding validity proofs also helps distinguish correctness from privacy. In November 2026, ZK withdrawals can take hours to a day, while optimistic withdrawals take roughly a week. These waits concern exits, not the speed of a sequencer's acknowledgment. Compare the complete path rather than one confirmation number.
Development moved from keeping individual payments off-chain toward publishing enough data for general execution elsewhere. Payment channels and Plasma preceded Ethereum's rollup-centric roadmap. Later changes expanded data capacity and made L2 maturity easier to discuss.
Joseph Poon and Thaddeus Dryja first circulated the Lightning white paper in 2015 and dated its current draft 14 January 2016. It described payments moving through connected channels, with value transfers occurring off-chain.
Poon and Buterin published Plasma on 11 August 2017. Its design posted Merkle roots, leaving users responsible for exiting when an operator withheld the underlying transaction data.
Blockstream's store took Lightning payments on Bitcoin mainnet with c-lightning in January 2018. Lightning Labs' lnd 0.4-beta followed on 15 March 2018 as its first mainnet release.
Buterin presented a rollup-centric roadmap in October 2020. It centered Ethereum scaling on separate execution, while transaction data and settlement remained connected to the original chain.
Arbitrum One opened to the public in August 2021. The network used the optimistic approach, which separated execution from the chain responsible for settling challenged results.
L2BEAT, now ETHBEAT, introduced Stages in June 2023. The framework added a maturity label, rather than treating the word rollup as a complete description of security.
Dencun activated EIP-4844 on 13 March 2024. Blobs gave Ethereum rollups a data publication path that was much cheaper than calldata, with an initial target of three.
Fusaka introduced PeerDAS on 3 December 2025. A blob-parameter-only fork raised the target to 10 on 9 December 2025, followed by 14 on 7 January 2026.
Security depends on permissions, available data and someone being able to act. The L2 label does not remove operational dependencies. These limits concern the design and its administration, rather than whether an application feels fast during normal use.
In ETHBEAT's 9 November 2026 snapshot, 22 rollups split into 12 at Stage 0, six at Stage 1 and four at Stage 2. Security councils can retain upgrade powers.
The Ronin sidechain bridge lost 173,600 ETH and 25.5 million USDC in March 2022 after five of nine validator keys were taken.
A BIS bulletin of July 2026 found that liquidity is split across many chains, so the same asset trades in separate pools on each.
In November 2026, Arbitrum One withdrawals wait 6 days 8 hours. ZK rollup exits wait for verified proofs, taking hours to a day.
A validium or optimium can verify results yet depend on outside data availability. If a committee withholds data, users can face frozen funds or restricted withdrawals.
Lightning requires inbound liquidity for incoming payments and monitoring for revoked states. A peer or watchtower must act within the delay, or cheating can stand.
We work on contracts and dApps deployed on layer 2s through our Ethereum development service, ZK rollup projects on established stacks, and custom chains. PixelPlex is a software engineering company founded in 2007 and building in blockchain since 2013. Before deploying on any layer 2, check its forced inclusion delay, its exit window and its Stage.
We built the Echo protocol as a chain with its own EchoRand consensus and sidechains to Ethereum and Bitcoin through a two-way peg. Its zero-knowledge rollup function executes transactions on Echo and posts validity proofs to layer 1.
Discuss your next layer 2 development decision
We can discuss choosing an existing network or building a system around your application's requirements.
Discuss your projectGet updates about blockchain, technologies and our company
We will process the personal data you provide in accordance with our Privacy policy. You can unsubscribe or change your preferences at any time by clicking the link in any email.Services
Blockchain
Blockchain consultingStrategy, architecture, feasibility
Crypto wallet developmentSelf-custody, MPC, cards on top
Cryptocurrency exchange developmentTrading venues, liquidity and protocols
Tokenization servicesRWA, funds, collateral
Crypto payments developmentAccept, settle and pay out in crypto
Smart contracts developmentDesign, build, audit
Blockchain integration dApp development DeFi development NFTs Tokens Web3 development All blockchain services All servicesNetworks
Canton developmentDaml apps, wallets, settlement, validators
Official Partner
Ethereum developmentSmart contracts, L2s and dApps
Solana developmentHigh-throughput programs and dApps
TON developmentMini apps, wallets, Telegram ecosystem
Official Partner
Hedera developmentHashgraph apps and token services
Exclusive Partner
Hyperledger developmentPermissioned networks for enterprises
Certified Partner
Cardano Algorand Stellar Avalanche Bitcoin Polygon All networks All servicesAI & ML
AI agents developmentConnect tools and automate multi-step workflows
RAG developmentFind answers across your company’s documents and systems
AI copilot developmentAdd useful assistance to your product or internal tools
LLM developmentCustom and fine-tuned models for your domain
Private AI deploymentRun models in your own environment
AI consultingUse cases, feasibility and a roadmap
Generative AI development Machine learning development Computer vision Chatbot development ChatGPT integration Deep learning development All AI services All servicesData & analytics
Data engineeringPipelines that keep your data reliable
Data platform developmentOne source of truth for analytics
Analytics & dashboards developmentBI your teams actually use
Predictive analyticsForecasts built on your own history
Data strategyWhere data can pay off first
Data visualizationDashboards and visual stories from your data
Data science development Development of data marts All data services All servicesSoftware
Product discoveryValidate scope and cost before you build
MVP developmentA launch-ready first release
Web application developmentPlatforms, portals and internal tools
Mobile app developmentiOS, Android and cross-platform
SaaS developmentMulti-tenant products that scale
Software integrationConnect the systems you already run
Infrastructure & cloud cost UI/UX design IT consulting QA & software testing R&D Serverless development All software services All servicesSecurity & compliance
Smart contract auditsFind vulnerabilities before launch
Transaction monitoring developmentIntercepta inside your systems and processes
Security assessmentsTest apps, infrastructure and processes
Data security servicesProtect sensitive data across its lifecycle
Cloud security servicesSecure cloud workloads and configurations
DAML smart contract auditReview DAML contracts before they go live
All security services All servicesSolutions
Products, APIs and platforms
New
Products and engineering to build and operate on Canton.
In production 55M API requests served in 8 months
Crypto APIBlockchain data for your applications.
Power in numbers 6 TB Data storage. 500K Tokens indexed
RWA platformA foundation for your asset tokenization platform.
Time to launch 2–4 weeks Branding & setup. 90% more cost-efficient
DocFlowBlockchain-based document management.
Use case & value 100% Secure workflows. Zero third-party risks.
OTC Hawk Arbitrage Bot Book a solution consultWork
450+ projects. 250+ clients. 3 unicorns in portfolio
The first non-custodial wallet built natively for Canton
TON ecosystemA series of improvements for The Open Network
Mining platformEnvironmentally friendly web platform for Bitcoin pool mining
Mantra DAOCommunity-governed DeFi: 60+ staking pools, $23M TVL
Canton TON Web3 & DeFi Digital Assets & Tokenization AI & Data Intelligence IoT & Edge Computing Consulting See all projectsIndustries
Trusted by startups, small businesses, and enterprises
Banking, payments, lending, insurance, capital markets
Real estate software developmentInvestment, property operations, tokenization
Retail & commerce software developmentCommerce, loyalty and fulfillment
Supply chain software developmentTraceability and connected operations
Healthcare software developmentPatient services, data and connected systems
Banking software developmentCore banking, digital channels, payments
Fitness Education Oil & Gas Insurance Construction TravelCompany
From R&D to launch, we bring any of your ideas to life
Engineering partner. Product builder. Since 2007.
Technical experience 19+ years All competencies on board. Scalable team.
AchievementsIndustry recognition & awards.
Delivering real value 4.9 Clutch ratings. 150+ awards
Contact usDiscuss your project.
Quick response 30 min Average reply time. Offices in 6 countries
BlogDive into the latest software trends.
Fresh insights 500+ Articles in blockchain, AI, big data, AR/VR
Team Careers Social responsibility Glossary