hedera hashgraph

How Hashgraph Works?


In the course of recent weeks, there has been a lot of discussion about a new technology known as Hashgraph. Up until now, many individuals have no clue what this is. As indicated by the Swirlds group, Hashgraph is a new consensus algorithm based on distributed ledger technology and its properties are:

  • Fast: 250,000+ Transactions Per Second
  • Fair: Consensus Time Stamping
  • Secure: Asynchronous Byzantine Fault Tolerant

The question is how exactly Hashgraph’s algorithm works. In this article, we take a look at Hashgraph’s algorithm. You can read my previous post to know more about Hashgraph’s launch event.

How it works?

Hashgraph’s whitepaper describes itself as an “algorithm for replicated state machines with guaranteed Byzantine fault tolerance”. Byzantine fault tolerance (BFT) is the dependability of a fault-tolerant computer system, particularly distributed computing systems, where components may fail and there is imperfect information on whether a component is failed. In a “Byzantine failure”, a component such as a server can inconsistently appear both failed and functioning to failure-detection systems, presenting different symptoms to different observers.

Hashgraph utilizes two extraordinary features to accomplish its quick, reasonable and secure transactions:

  • Gossip about gossip
  • Virtual Voting.

Gossip about gossip: Hashgraph utilizes a gossip protocol. This implies a part, for example, Alice will pick another member indiscriminately, for example, Bob, and after that Alice will reveal to Bob the majority of the data she knows up until now. Alice then repeats with a different random member. Bob repeatedly does the same, and all other members do the same. In this manner, if a single member becomes aware of new information, it will spread exponentially fast through the community until every member is aware of it.

Consider an example for this:

The above figure is a hashgraph. It develops upward after some time. Each member keeps a duplicate of it in memory. In this case, there are four individuals (full nodes). They are in network. The individuals are Alice, Bob, Carol, Dave, and are represented with 4 lines named A, B, C, D. Each member begins by starting an event, which is a little information structure in memory, and which is shown by a gray circle here. After this the gossip protocol starts.

Bob randomly calls Dave. When they connected with each other on internet, Bob sent Dave each and every event/information he knew that Dave did not yet know. In this example you can see from above figure that it was just one event: the one that Bob had created at the start. Dave creates a new event. This is the new circle, which has lines going straight down to his own last event, and diagonally down to Bob’s last event. Thus, the graph of events forms a record of how the members have communicated.

Now, if Dave wants to randomly interact with Bob again then Bob now has three events, the first event of both Bob and Dave and the new event that was created by Dave. This goes on and on and the hashgraph will look like below figure:

Virtual Voting: If exchange/transaction has 2/3 of the hub in the system as witness, at that point it is viewed as legitimate. Keep in mind, hashgraph was intended to be Byzantine Fault Tolerant, thus this calculation enables the framework to work regardless of whether 33% of the hubs turn Byzantine.

Hashgraph is turned out to be completely asynchronous Byzantine. This implies it makes no suppositions about how quick messages are s ent over the web and this makes it strong against DDoS assaults, botnets, and firewalls. The Hashgraph calculation works without expecting to utilize the Proof of Work or Leader frameworks, and can likewise deliver low-expenses and high performance without failure. Hashgraph gets rid of the requirement for broad calculation and vitality utilization and enhances the execution insights of the Bitcoin network. Bitcoin works at a most extreme of 7 transactions per second. While Hashgraph is just restricted in connection to data transfer capacity and takes into account more than 250,000 transactions per second.


Well, it is too soon to state that Hashgraph is better than Blockchain. What can be said however is that the innovation behind Hashgraph is fascinating. It stays to be perceived how powerful people in general, non-permissioned adaptation of this innovation resembles.