Directed Acyclic Graphs (DAGs) have been common in the realm of computer science for a while now, but it was only in 2015 that the first proposal was put forth to cross it with blockchain technology.
Sergio Demian Lerner first suggested the idea when working on Dagcoin – a project which failed to come to fruition, but which ultimately opened the door for other developers to build upon the DAG concept.
DAGs are not actually blockchains, since there is no sequential chain which links all the blocks. In fact, there are no blocks at all, and the transactions which occur on individual nodes do not need to be synchronized with any other. By allowing transactions to occur without the confirmation of the entire network, the time it takes to receive a confirmation is greatly reduced.
Reduced transaction times are just one of the benefits of DAGs. Bitcoin’s high fees and long delays during the boom at the start of the year acted as a strong reminder that blockchains aren’t the only game in town.
How do DAGs Work?
On a DAG there is no need for miner rewards, and no transaction fees for the end user. Transactions are confirmed via a process of ‘paying it forward’, where each new user confirms two previous transactions for their own transaction to be processed.
Every transaction acts as its own block, and these can be stored in disparate locations, separated on multiple devices, before eventually syncing up with a node somewhere on the graph. The syncing process updates the ledger with the final transaction details (i.e. balances) of all the interactions which took place among the addresses on the graph.
Scalability isn’t an issue with DAGs since the process of confirming the two previous transactions can only be sped up by an increased user base.
Who’s Using DAGs?
The most well-known adopter of DAG technology is IOTA, who have utilized it in a set-up they refer to as the Tangle.
The Tangle is the web that makes up IOTA’s network of users – who act as both transactors and verifiers at the same time.
The IOTA whitepaper, published in 2016, describes the process:
“…to issue a transaction, users must work to approve other transactions, therefore contributing to the network’s security. It is assumed that the nodes check if the approved transactions are not conflicting and do not approve conflicting transactions. As a transaction gets more and more approvals, it becomes more accepted by the system; in other words, it will be more difficult to make the system accept a double-spending transaction.”
The lightweight and decentralized nature of the Tangle makes it a strong candidate for Internet of Things (IoT) utilization. Because of the ability to introduce new data into the Tangle via independent nodes, IOTA also stands to be well utilized as an oracle tool, where it can rapidly disseminate the data required for prediction markets to function.
Byteball Bytes and Hshare
Byteball Bytes (GBYTE) have been around for as long as IOTA, and although less well-known, provide many of the same services as IOTA. They have also implemented atomic change capabilities, multi-signature practices, conditional payments (arbitrated by smart contracts) and on-chain oracles.
Hshare, sometimes known as Hcash, also uses DAG structures but to a different degree. Hshare uses a hybrid of both blockchain and DAG technology, and the project is set up to act as a future bridge between blockchain and blockless systems.
Downsides
As glorious as some of the DAG’s features may seem, every technology comes with its weak points, and directed acyclic graphs are no different.
While many DAG-coins claim to be quantum-resistant, there remains some doubt over their ability to survive a 33% attack – where someone with enough computing resources could essentially take control of the entire network.
At the same time there are questions over the ability for DAGs to be truly decentralized, since, in the case of IOTA specifically, the developers must run their own coordinator node at all times in order to protect the network from the aforementioned 33% attack.
The majority of Bitcoin is also ‘controlled’ by a small number of large mining pools, so the dangers of centralization exist for blockchains as well as DAGs.
