The Tangle is the underlying consensus data structure of IOTA, described in a 2016 whitepaper (revised 2018) by probabilist Serguei Popov. Unlike blockchain systems where transactions are grouped into sequential blocks by miners, the Tangle is a Directed Acyclic Graph (DAG): each new transaction directly references and approves two previous transactions, building a web of approvals rather than a linear chain.

The design targets IoT (Internet of Things) and M2M (machine-to-machine) payments: feeless microtransactions between sensors, devices, and automated systems — a use case where even sub-cent transaction fees are prohibitive.

> Whitepaper (PDF): iota1_4_3.pdf

---

Publication and Context

IOTA launched in 2016 founded by David Sønstebø, Sergey Ivancheglo (Come-from-Beyond), Dominik Schiener, and Serguei Popov. The project raised ~$500K in a public token sale in autumn 2015 and launched mainnet in July 2016. MIOTA peaked at a $14.5B market cap in early 2018, making it one of the top 5 cryptocurrencies by market cap at the time.

The core thesis: blockchains are the wrong data structure for IoT. IoT devices perform millions of tiny value transfers (e.g., a sensor selling weather data, a car paying for parking by the second). These require:

• Zero fees: A 1-cent fee on a 0.01-cent payment is absurd
• High throughput: Millions of devices simultaneously
• Lightweight participation: Each device should be able to validate its own transactions

---

The Tangle: DAG Structure

In the Tangle, each transaction (site) must reference exactly two prior transactions (tips):

Transaction A  Transaction B
                     /
                    /
        Transaction C  (references A and B)

By referencing A and B, Transaction C approves them: it has verified their validity and included them in its confirmation path. This is the fundamental innovation: every user is simultaneously a validator.

Tip Selection

• Each step moves forward along the DAG proportional to cumulative weight (approval count from descendant transactions)
• Transactions with more approvals are more likely to be selected, creating natural convergence
• Tips receive higher confidence as more transactions approve them

Cumulative Weight

Proof-of-Work Per Transaction

---

Coordinator (The Centralization Problem)

The whitepaper assumes the network has a sufficiently large number of honest participants. In practice, IOTA’s early network was too small and sparse for the tip selection algorithm to converge on canonical confirmations.

The Coordinator (COO): The IOTA Foundation ran a special node called the Coordinator that periodically issued milestone transactions. Only transactions referenced (directly or indirectly) by a milestone were considered confirmed.

This was an explicit centralization compromise — the IOTA Foundation could halt the network by stopping the Coordinator, and they did so during the February 2020 hack (of the Trinity wallet) where they halted the Coordinator for 11 days to prevent theft of user funds.

Project Chrysalis / IOTA 2.0 (Coordicide): The multi-year effort to remove the Coordinator involved:

• Mana: A reputation system based on past token activity, used for Sybil resistance
• Cellular Consensus / Shimmer: A gossip-based rapid finality mechanism replacing tip selection
• IOTA 2.0: Launched on testnet (Shimmer EVM, Stardust protocol); Coordinator removal achieved in testnets but as of 2024 had not fully launched on mainnet

---

Key Technical Properties

Property	Value
Data structure	DAG (not blockchain)
Transaction fees	Zero
Spam prevention	Per-transaction PoW
Finality mechanism	Tip selection + Coordinator (legacy)
Coordinator removal	Planned (IOTA 2.0 / Shimmer)
Target use case	IoT, M2M, microtransactions
Token symbol	MIOTA (1 million IOTA)
Total supply	2,779,530,283,277,761 IOTA (fixed, all pre-mined)

---

Reality Check

IOTA articulated a genuine use case (feeless IoT payments) and an innovative data structure (DAG). In practice:

• The Coordinator was a critical failure of the whitepaper promise. IOTA marketed itself as decentralized while operating a centralized coordinator for most of its existence.
• The February 2020 Trinity wallet hack allowed attackers to steal ~$1.6M in MIOTA by compromising the official wallet’s third-party API dependency. The Coordinator was stopped for 11 days — demonstrating that IOTA’s “decentralized” network was controlled by a central switch.
• Coordicide delays: Promised repeatedly since 2019, the Coordinator-free IOTA 2.0 repeatedly missed target dates, eroding community trust.
• IoT adoption: Despite partnerships (Volkswagen, Bosch), real-world IoT deployments using IOTA remained minimal.
• Ternary abandonment: IOTA originally used balanced ternary (base-3) arithmetic to optimize for theoretical ternary hardware that did not exist — this created compatibility problems and the ternary design was eventually abandoned.

---

Legacy

IOTA pioneered DAG-based distributed ledger design, directly inspiring later projects (IOTA → Nano’s block-lattice, Avalanche’s DAG consensus, Kaspa’s GHOSTDAG). The Tangle whitepaper’s mathematical analysis of DAG tip selection and cumulative weight remains cited in academic literature. The IoT payment thesis, while unproven at scale, remains a valid research area.

---

Related Terms

• DAG
• Proof of Work
• Feeless Transactions
• Nano Whitepaper

---

Research

• Popov, S. (2018). The Tangle (v1.4.3). IOTA Foundation.

— Primary whitepaper; mathematical analysis of DAG tip selection, cumulative weight, and convergence properties under various honest/adversarial assumptions.

• Ferraro, P., King, C., & Shorten, R. (2018). IOTA-Based Directed Acyclic Graphs Without the Coordinator. IEEE Access.

— Independent academic analysis of the Tangle’s convergence properties without a coordinator; contributes to the Coordicide research.

• Silvano, W.F., & Marcelino, R. (2020). Iota Tangle: A Cryptocurrency to Communicate Internet-of-Things Data. Future Generation Computer Systems 112.

— Survey of IOTA’s architecture and IoT applicability; discusses feeless microtransaction properties and performance benchmarks.