At ProvenDB, our mission is to provide a platform with the functionality, performance, and economics of a traditional database system together with the immutability and trust offered by public blockchains. From a business perspective, we aim to be blockchain agnostic – supporting as many public chains as possible and even a few private chains such as Hyperledger. Nevertheless, we have needed to implement a default Blockchain – particularly for free accounts - and we are often asked by potential customers which chain we recommend.
Until recently, our default blockchain has been Ethereum. Ethereum has tremendous mindshare, especially amongst leading-edge blockchain developers, and – at least until recently – was economical, at least in comparison to Bitcoin.
However, we recently decided to switch our default anchoring protocol to Hedera. In this blog post, I want briefly to explain our decision-making process and share our experiences with Hedera Hashgraph.
One of the advantages of ProvenDB is that we can anchor large amounts of data to a Blockchain at massively lower costs than a DIY approach. When data in a ProvenDB database is anchored, we create a Merkle tree which consolidates potentially hundreds of thousands of database records into a single root hash which is anchored to a chain in a single transaction. Furthermore, multiple customers can “piggyback” on this transaction. So a single transaction can support numerous customers and masses of data.
Therefore, the cost of doing business for us depends on how frequently we send Blockchain transactions. If we want to offer customers Blockchain proofs with 2 minutes latency, we need to be prepared to issue 720 proofs per day.
When Ethereum transaction fees were in the vicinity of about 20 cents per transaction, this was no big deal. However, over the past few months, we have seen Ethereum transaction fees spike up to $50 or more. Suddenly, we spent tens of thousands of dollars per day, primarily to support free accounts!
To add insult to injury, confirmation times for Ethereum transactions were also increasing. Ethereum transaction confirmations work on a bidding scheme. An application can pay a higher transaction fee for faster confirmations. However, you must estimate your transaction fee in advance. If you calculate incorrectly, your transaction may languish on the chain for minutes or hours before being confirmed.
This spike in price reflected a spike in utilization brought on by the boom in crypto trading and the adoptions of NFTs. But from our point of view, Ethereum had become both expensive and unreliable.
We had supported Hedera Hashgraph as a first-class anchoring mechanism for quite some time and using it as our preferred chain internally for over a year. Once we hit these Ethereum problems, the solution was obvious: we switched our default anchoring mechanism to Hedera Hashgraph and made Ethereum proofs optionally available for paid accounts only.
The Hedera team has built a technologically sophisticated and impressive platform, and we are true fans of what they have created. But they also made an incredibly wise business decision when they priced their service. They decided that their offering should be economical and predictable in price. Unlike virtually any other public ledger technology, Hedera transactions are always fixed-price in US dollars. A Hedera Consensus Service transaction, given our larger message size, always costs us $0.10, regardless of the demand on the network and regardless of the relative price of the Hedera cryptocurrency HBAR.
Additionally, Hedera transactions are very fast: typically confirming within a few seconds, compared with the minutes or hours sometimes experienced on other chains.
I hope that explains why we have settled on Hedera Hashgraph as our primary anchoring technology. We are still big fans of Ethereum and were looking forward to the Ethereum 2.0 rollout that may address some of the issues we have encountered. But for now, the choice seems clear: Hedera Hashgraph offers more economical, performant, and predictable service than any of the alternatives.