DAG (Directed Acyclic Graph)

What Is a DAG (Directed Acyclic Graph)?

Picture a family tree. Every person points back to their parents, the arrows only go one way in time, and nobody can be their own ancestor. That is a directed acyclic graph: directed because each link has a direction, acyclic because you can never follow the arrows in a circle, and a graph because it is a network of points and connections.

Computer scientists used DAGs long before crypto. Git stores code history as one, spreadsheets use one to decide which cell to recalculate first, and build tools use one to order tasks. A blockchain is actually the simplest possible DAG, a single line where each block points to exactly one parent.

DAG ledgers relax that single-parent rule. A new entry can reference several earlier ones, and several entries can be added at the same moment. That changes how a network reaches agreement, which is where the trade-offs begin.

How Does a DAG Work in Crypto?

With the basic shape in mind, look at why the single line limits a blockchain. Bitcoin adds one block roughly every 10 minutes, and if two miners find a block at the same time, the network keeps one and discards the other as an orphan. That discarded work is wasted, so the block time must stay long enough for new blocks to spread across the world before the next one appears.

A DAG ledger keeps the parallel entries instead of throwing them away. In IOTA’s Tangle, launched in 2016, every new transaction must approve two earlier unconfirmed transactions by checking that they do not conflict.

Suppose Alice sends 100 tokens to Bob. Her wallet picks two recent transactions, verifies them, links to them and broadcasts the payment. Over the next seconds, dozens of later transactions link to hers directly or indirectly, and each new reference adds weight. Once that accumulated approval passes the network’s threshold, her payment counts as confirmed.

The hard part is conflict. If Alice tried to spend the same 100 tokens twice in two branches, the graph would contain both, and the network needs a rule to pick one. Each DAG project solves this double spending problem with its own ordering algorithm. Some use voting among nodes, some weight branches by accumulated work, and others give a trusted coordinator the final say.

Types of DAG-Based Networks

• Transaction DAGs such as IOTA’s Tangle have no blocks at all. Each transaction is a node in the graph and confirms earlier ones.
• Gossip-based DAGs such as Hashgraph, used by Hedera, record who told whom about which transaction, and nodes calculate a fair order from that shared history.
• Block-lattice designs such as Nano give every account its own mini-chain, so only the sender and receiver’s chains need updating.
• BlockDAGs such as Kaspa keep mining blocks with proof of work but allow several blocks to reference each other in parallel, then order them with the GHOSTDAG protocol.

DAG vs. Blockchain

Why Is a DAG Important for Traders?

A DAG is a bet on scalability. Projects built on one promise high transactions per second and low fees, which matters for micropayments, machine-to-machine data and trading apps that settle constantly. When you evaluate such a token, the data structure tells you what the team is optimising for and which problem it believes blockchains cannot solve.

The weakness usually hides in the ordering rule. Many DAG networks launched with a central component to prevent double spends, and that component became a single point of failure. In February 2020 the IOTA Foundation shut down its Coordinator node for almost four weeks after attackers stole funds through a wallet vulnerability, which stopped transactions across the entire network. For a trader, a pause like that means you cannot move coins to an exchange during a selloff.

Governance is the second concern. Hedera’s ledger is run by a council of large companies rather than an open set of validators, and several other DAGs rely on foundation-operated nodes. Always check who can confirm transactions today, not just what the whitepaper promises for the future.

Key Takeaways

• A directed acyclic graph is a network of records connected by one-way links with no loops, and a blockchain is its simplest form with a single parent per block.
• DAG ledgers let many transactions or blocks be added in parallel, with each new entry confirming earlier ones, instead of discarding simultaneous blocks as orphans.
• Because parallel branches can conflict, every DAG network needs an extra ordering rule to prevent double spending, and that rule defines its security.
• The main designs are transaction DAGs, gossip-based hashgraphs, block-lattices and proof-of-work blockDAGs.
• Higher throughput often comes with centralised coordinators or permissioned governance, which can halt the network or concentrate control.