Think of blockchain as a staircase and DAG as a tree.
Although often mistakenly placed under the same umbrella with “blockchain technology”, DAG isn’t technically similar to blockchain at all.
Think of a standard blockchain like a staircase. Each step you add to the staircase is like a new block added to the blockchain, extending it in a single direction.
Synchronising blocks one at a time, like stacking a staircase, guarantees that transaction records are the same across all the nodes in a decentralised network. If a malicious actor tries to manipulate the blockchain and create a fake transaction, the nodes can now easily determine that it’s a fake and exclude it from the block.
Blockchain technology, therefore, provides a way to maintain consensus, security and trust across all nodes, without a central authority.
This is a major breakthrough.
To guarantee a good quality staircase, building steps takes time. You may even ask others to quality-check your work just to be sure. Generating new blocks is also bound by time, because the chances of fake transactions slipping into a block reduce when nodes have more time to process transactions and do block confirmations. Bitcoin’s block time is 10 minutes, Ethereum’s 16 seconds.
To make sure the staircase is safe, it needs regular maintenance. If it’s a popular staircase, perhaps you set up a small toll fee so that whoever uses the staircase contributes to maintenance. Similarly, blockchains have transaction fees, because the possibility of getting paid well motivates nodes to do their best at maintaining block records and keeping them identical across the network. Fees also make it expensive for a malicious actor to cripple the network with a flood of fake transactions.
Blocks, block times and transaction fees are therefore some of the reasons why blockchains are more secure than current centralised database technology.
However, some of the biggest factors holding it back from being used more widely are the ability to scale and settle transactions fast. Processing blocks one by one and under a strict time limit slows a blockchain down significantly. On average, Bitcoin is able to process only 7 transactions per minute. And this is why Ethereum has trouble running more than a few large dAPPS concurrently on the network — during times of high traffic, slow transaction rates create a tight bottleneck and a backlog of transactions that can take days to catch up. When this happens, transaction fees escalate too, because senders with important transactions compete to get theirs broadcasted and confirmed faster, making it unsustainable to perform smaller transactions, like paying for a coffee.
This is exactly what happened in the December Bull Run in 2017. Bitcoins were sent back and forth with such volume that the fees were more than 80 USD and the backlog 3 days long. Around the same time, CryptoKitties, a popular dAPP trading digital cryptographic kittens, “broke” Ethereum by taking up 25% of the network. By contract, blockchain’s biggest financial processing competitor Visa is able to process 25,000 transactions per second.
So, decentralised technology with synchronous blocks and limited block times are arguably not yet scalable enough for large scale use. To address this deficiency in speed and scalability, DAG does away with blocks completely.
Think of DAG like a tree, where transactions are processed in a network of interlocking branches that grow outwards in multiple directions. In DAG, each transaction only needs to confirm the validity of a previous transaction to become valid itself. If a stand-alone transaction is found without a valid transaction preceding it, the nodes can easily identify it and exclude it from the records.
A single valid transaction can confirm many transactions in succession, growing transaction output like fractal branches on a tree. In fact, the more transactions are processed, the more it becomes possible to process more. So when a DAG system is experiencing high traffic, the very presence of more transactions makes it possible for DAG to meet the increased demand.