WHAT IS DIRECTED ACYLIC GRAPH (DAG)?

Most of the time, in our daily conversations, DLT and Blockchain are used interchangeably. This can be attributed to the popularity that Blockchain has gained over the years with Bitcoin and the like. Distributed Ledger Technology, like the name suggests, is a decentralized ledger of transactions or contracts that is accessible to all the participants across different geographies. This decentralization eliminates the need for a central authority to keep track and validate the transactions within the chain. To explain this, Consider security personnel of a gated community, who takes the details of every person entering through the gate in a book(ledger). The case above is a centralized database; if someone wants to manipulate the data, all they need to do is to make changes in that particular ledger. Now, imagine an app, which requires each guest to fill in the details, and the details will be sent as a message to all the occupants. In this case, every occupant will have data in their mobiles in the form of a text message. If the person wants to manipulate the data, he needs to do that to all the databases(mobiles), which makes it complicated if not impracticable. 

Immutability of data, Instant updating of all the records, and decentralized database makes DLT, a plausible innovative technology in the areas of finance, operations, supply chain, etc.

Few popularly known types of DLTs include Blockchain, Hash Graph, DAG, Holochain so on so forth. Blockchain is the most commonly used and talked about of all. In a blockchain, every time a transaction occurs, it is verified, the blocks store information about the transaction like time, date, and each block is then given a hash, a unique identification. Every participant will have a copy of the Blockchain, thereby reducing the security concerns. Its structure might seem similar to our linked list structure, linearity on the face of it. But it is more complicated than the traditional list-like structure. One major shortcoming of this entire structure is the low number of transactions per second. If we are going to scale it up for much more extensive and diversified uses, this set back needs to be addressed. 

 

DAG EXPLAINED

Directed Acyclic Graph is one of the highly scalable alternatives to Blockchain technology. In layperson’s terms, this structure could be viewed as a tree structure, which will address the scalability issues in a list type arrangement. IOTA with Tangle Protocol and Hashgraph with Gossip Protocol are few examples of DAG that are in usage. DAG comprises of a network of nodes confirming transactions. Each node will perform two functions, validating transactions and representing transactions themselves. Two tasks at hand are initiation and validation; any node can initiate a transaction, but for a node to get verified, it needs to perform at least two validations. The graph will not be allowed to grow further in the case of invalid descendants.

• Highest Weighted node in the above graph
• Validated two transactions(min limit), hence valid node in the above graph
• Validated less than two transactions, need to validate more to become a valid node  in the above graph

 

The width of the graph is the number of nodes that are the same depth. Depth implies the maximum length of the path from a node to its ascendants. In the initial phase of the graph, we might not have enough descendants to validate, in which case DAG goes for non-terminal parent selection. However, this might lead to an issue where there is a rapid increase in the width of the graph, to avoid this transaction time out can be used. In the case of an invalid node, descendants will also get affected, which is quite not right. Parent reselection can be used to avoid this situation and invalidating the node so that ancestors don’t get altered.

The graph becomes more and more robust with increasing nodes/transactions. DAG requires no miners to confirm the authenticity of each transaction. Two ascendants transaction validations confirm the validity of a descendant node transaction, human intervention becomes trivial, resulting in a vastly accelerated process. This implies transactions happen almost instantaneously due to reduced human intervention. In case of Blockchain significant part of cost involve the miner fee, given DAG reduces miner interaction, this cost is notably reduced. This low-cost structure leads to the possibility of nano transactions.

More significant, the depth of the particular node, the more the weightage attributed to it, and also non-terminal node selection in case of low volume (nodes/ transactions) make DAG vulnerable to attacks. However, this might not be the issue as applications of large scale use DAG.

Recent days, few papers have been published on DAG in Distributed ledger technology applications on smart grid networks, IoT, Cryptocurrencies, Smart mobilities, Smart Cities etc., Applications that require high scalability and speed along with features of DLT can opt DAG as one of the solutions. However, like other DLT processes, DAG is in its nascent stage, and its potential needs to be investigated more for its applications other than cryptocurrency. Future, where minimum transaction fee and high functioning chains are possible, doesn’t seem farfetched with DAG at the helm. This means all our day to day transactions (micro and nano) can be carried out without hefty fees, unlike other DLTs with improved security.