Blockchains are often seen as the preserve of the public, as open source, permissionless systems that allow transactions to be sent privately, but with all transactions publicly available. However, there is a subset of blockchains that eschew decentralization and the open-source approach in favor of speed and privacy – enterprise blockchains.
Enterprise blockchains are used by businesses who want the benefits of blockchain technology but without the issues of speed and scaling that plague public blockchains likes Bitcoin and Ethereum. They also don’t want the transactions to be viewable by Joe public, only by those whom they choose to give permission.
This has given rise to a number of blockchains that fulfil this brief, and in this piece we run through three of the most used and best known enterprise blockchain to see which is the best of the bunch.
Why do enterprises want blockchains?
Blockchains help businesses because they open up data to all parties involved in a process, resulting in fewer errors and communication breakdowns. This, in theory, leads to systems moving quicker and smoother, which is to everyone’s benefit.
Using logistics as an example, just one international goods shipment can involve multiple different types of cargo being transported across the world together, requiring clearance through numerous countries, each with its own way of doing things. Throw in language differences and it’s easy to see why the logistics industry, which still relies heavily on paper documents, is crying out for a modern solution.
Blockchain technology means that data on the shipment is fed to a central, immutable ledger that all parties can view, meaning that all stakeholders can track every aspect of the shipment whenever they like.
The benefit of this is that the element of trust is removed from the process – there is no need to rely on the word of someone halfway across the world that your shipment is where it is supposed to be. The data doesn’t lie. Blockchain technology offers trustless solutions that are already being used in various sectors around the world.
What do businesses need from a blockchain?
When it comes to which blockchain offering is the best, we of course need to look at what businesses want from such a system – or more accurately, what they don’t want. Key to this is the premise of decentralization. Decentralization means that no single party is in control of the blockchain.
Instead, a series of independent nodes spread over the world verify all transactions, which are all publicly viewable. Clearly, unless they have a specific ethos, this is not something that companies generally seek – they want to control access to the data and how it is verified. Also, decentralized blockchains almost always tend to be slower than centralized ones, and their scaling has so far turned out to be less than impressive.
Enterprises don’t want to be told that they have outgrown the capabilities of the blockchain – they want it to be able to expand with them. Allied to this, businesses want blockchains that are interoperable with other financial systems and, in the future, other blockchains.
The three blockchains we will examine now are not ones that most cryptocurrency users will be familiar with, but they are blockchains that have been doing the rounds in enterprise circles for over six years and are already offering some pretty stiff competition to the status quo.
Quorum is essentially an enterprise version of the Ethereum blockchain but with so many differences that Vitalik Buterin would find it hard to recognize his creation. Quorum dispenses with Ethereum’s famed decentralization, replacing it with a more privacy-focused model which allows it to beef up throughput by around 10x compared to its decentralized cousin.
Not only are the transactions near instant they are also free, something that regular Ethereum users would consider a fantasy scenario.
Quorum was forked from the original Ethereum model in 2016 by JPMorgan who began to develop it as a blockchain on which the bank could eventually migrate its infrastructure, saving millions of dollars in costs by doing everything on the blockchain.
JPMorgan has since launched its own token on the Quorum blockchain, JPMCoin, and operates the Interbank Information Network which currently boasts 350 members, who all use Quorum and JPMCoin to do business on a daily basis.
Quorum itself features four key differences from Ethereum that make it suitable for enterprise use:
Permission management – Quorum is what’s known as a ‘permissioned blockchain’, which simply means you need permission to be able to use it. A designated authority must grant permission, meaning that transactions cannot be carried out on the blockchain without prior approval.
Privacy – transaction details are recorded in the same way as with the Ethereum blockchain, but the details are only open to approved parties. This is in a large way down to Constellation, a mechanism that encrypts specific messages in an enclave, allowing transactions to be carried out in cryptographically secure manner.
Voting – Protocol voting is managed by QuorumChain, the native voting consensus mechanism which assigns voting rights to parties and tracks the status of all the voting nodes. It operates as a majority voting protocol, meaning that a transaction is completed if it receives a majority of votes.
Performance – due to its development and its centralized nature, Quorum can boast over 100 transactions per second (tps), more than 10x that of regular flavor Ethereum. Such throughout is of course critical to enterprises who are used to transactions reaching their destinations at the click of a mouse. With further development in the pipeline there is a likelihood that Quorum will increase throughput further over the years.
As well as these key benefits, enterprises can also create their own blockchain networks on Quorum from scratch, managing various internal and external elements to make it unique to their business.
While Quorum has some important benefits it also has some drawbacks, the major one being scaling. When Quorum was forked from Ethereum in 2016 there wasn’t much competition around and no real thought had been given to scaling the blockchain.
This has left Quorum with a limited block size that, despite its supposed 100 tps throughput, may well see transaction queues at busy times. In fact, web developer Scandiweb tested Quorum in 2018 and found “issues with transaction pool processing, which can either be painfully slow or infinite”, although it concluded that the advantages of Quorum outweighed these negatives.
Testing has revealed that Quorum more or less meets its performance goals. A 1,000-transaction test by Menapay in 2018 revealed throughput of between 65-152 tps depending on the demands, while the Central Bank of South Africa performed a “realistic” settlement pilot in June 2020 which saw 70,000 transactions processed in two hours, with anonymity preserved, at an average rate of 1-2 seconds per transaction. This equates to a tps of 30-60.
When it comes to interoperability, this is where ConsenSys, Quorum’s owner since August 2020, comes in. ConsenSys bought Quorum with the idea of making private blockchains like Quorum and public blockchains like Bitcoin and Ethereum communicate easier.
This is a big part of the move to Web 3.0, which is a vision of a decentralized internet. While Quorum may not be there yet, with ConsenSys making interoperability its number one priority it might not be long before Quorum is chatting happily to its common ancestor Ethereum.
Quorum ticks many of the boxes needed by enterprises wanting to use a blockchain, although the throughput, constrained by the Ethereum model, may be an issue for some. This doesn’t seem to have put off the members of the Interbank Information Network however, who happily use it on a daily basis.
Hyperledger Fabric is one of a cluster of iterations under the Hyperledger name, a project which was founded by Linux in 2015 and has seen contributions by several tech giants including IBM, Intel, and SAP Ariba. IBM is now the chief developer and user of Hyperledger Fabric, which it calls the “de facto standard for enterprise blockchain platforms” and which underpins the wildly successful IBM Blockchain protocol.
Hyperledger Fabric is a modular, general-purpose blockchain that offers unique identity management and access control features, making it suitable for a variety of applications across multiple industries such from supply chain to global finance.
Powered by Hyperledger Fabric, IBM Blockchain operates the 300-member strong TradeLens supply chain ecosystem, which is one of the most successful exponents of the enterprise blockchain movement in the world. As well as powering IBM Blockchain, Hyperledger Fabric is used by the likes of PayPal, MasterCard, and Amazon, giving it a great pedigree when it comes to considering the best blockchain for enterprises.
This was emphasized in 2019 when Hyperledger revealed that half of the companies in Forbes’ Blockchain 50 list used Hyperledger technologies.
Technically speaking, Hyperledger Fabric makes bold claims, chief among them that its blockchain can push through a staggering 20,000 tps. This is still less than halfway towards Hyperledger developers’ ultimate goal of 50,000 tps. However, this assumes ideal conditions which is never the case in the real world and doesn’t account for network load.
Unfortunately, data for real world tests is hard to come by and results are sporadic, with tps figures ranging from 300 to 11,000 and transaction latency ranging from 1-10 seconds. Hyperledger itself notes that prior to its 1.1 release “performance was not great” and that even now “we can actually scale…if we are willing to tolerate additional latency”, which means that the top speeds Hyperledger advertises doesn’t take into account transaction times which could make it impracticable for many enterprises.
Indeed, this disparity over real world speeds versus theoretical speeds is the chief criticism of Hyperledger Fabric, with several researchers finding that IBM’s grand claims simply don’t add up. Indeed, some have gone further and suggested that Hyperledger Fabric isn’t even a blockchain at all.
Regardless of whether Hyperledger Fabric really can reach those headline figures on a consistent basis and with a good latency, it’s clear that even at lower speeds it is a very fast blockchain.
When it comes to scaling, someone somewhere at some time started a rumor that Hyperledger can only scale to a set number of nodes, limiting its ability to scale dramatically. Hyperledger confirms that this is, indeed, just a rumor, pointing to tests as far back as 2019 in which it claimed its throughput reached 13,000 tps with 128 peers.
To put this into context however, in 2019 SWIFT said that the company itself would need more than 100,000 nodes on Hyperledger Fabric to properly operate Nostro accounts with 34 banks.
Deploying extra nodes is not simply done with the click of a mouse – they are complex to configure, and when you’re talking hundreds or thousands per client there has to be a point where the reward isn’t worth the hassle.
Like Quorum, Hyperledger Fabric allows enterprises to mould a blockchain network around their needs, assigning the number of nodes and their roles up front to make sure that the network is right for them.
Whether the plethora of enterprises that have signed up with IBM Blockchain are regretting their decision or whether they are getting the speeds and throughput they promised we will never know, but the fact that IBM Blockchain and Hyperledger Fabric has achieved such stellar market penetration shows that they are clearly doing something right.
When it comes to interoperability, Hyperledger Fabric is great when it comes to connecting to the other versions of Hyperledger but not so great in terms of connecting with public blockchains. It is, however, working on solving this with Hyperledger Cactus, a plugin designed to allow Hyperledger Fabric to do just that.
While it certainly talks a good talk, Hyperledger is yet to conduct any public tests with Cactus, which shows just how early it is in the interoperability stakes. Should the tests be unsuccessful, it might be some time before Hyperledger Fabric can work with other blockchains, a fact that adds to the element of uncertainty around Hyperledger’s lofty goals for the project.
Corda development was initiated in 2015 by a consortium of nine major banks including Bank of America Merrill Lynch, UBS, Credit Suisse, and Wells Fargo. It is unsurprising, then, to know that Corda was aimed from the get-go at the institutional banking sector.
Corda goes about smart contract execution in a different way to traditional blockchains. On regular blockchains, all the nodes store their own copy of the blockchain and all transactions made on it right back to the first ever transaction.
Corda approaches things differently, in that only the parties involved in the execution of a smart contract have access to it, as well as any legal or regulatory bodies that the parties agree should see the information. This ensures that transactional details are permissioned at every stage, increasing privacy and offering companies complete control of their transactional data, both current and historical.
Corda also handles transactions differently than Hyperledger Fabric and Quorum. Corda transactions are based on a UTXO (unspent transaction output) model instead of an ‘account state’ model which allows for concurrent processing compared to the linear processing used by the other two blockchains: assets are split into multiple parts and transacted concurrently rather than a single asset waiting for its turn to be sent.
This doesn’t work for non-fungible assets (e.g. NFTs) however which cannot be split, negating the benefits of the UTXO model in these instances.
Corda also collects transaction signatures in a more efficient way than its competitors. Whereas Quorum and Hyperledger Fabric typically require applications themselves to coordinate signature collection, Corda uses ‘flows’ to collect signatures. Flows automate the various steps needed for network participants to specify exactly what information needs to be sent to which counterparties and in what order.
This means that any extraneous information is stripped out and only those who need to know information in order to update and synchronize ledgers are informed. As a result, transaction confirmation times are as quick as they can possibly be, even when a blockchain is congested.
When it comes to speed, Corda is cagey about its blockchain’s throughput and transaction times, saying that the metric is often misinterpreted and incorrectly used as a barometer of performance. They are of course right in many ways as we have explained, but nevertheless it does give us an indication of a blockchain’s prowess.
In February 2018 Corda offered two figures as a guide to its speed and throughput – 1,678 tps and 170 tps. The higher figure was a theoretical throughput across one node, whereas the lower figure was a full end-to-end transaction between two parties.
This doesn’t exactly blow the others out of the water, putting it firmly in Quorum’s camp, but a test by The Depository Trust and Clearing Corporation (DTCC) 10 months later found that Corda completed “6,300 trades per second for five continuous hours” across 170 nodes. This is much more impressive and echoes Corda’s desire “to keep pushing the 1,678 and 170 TPS figures higher”.
Corda can also boast its fair share of stellar clients. Corda creator R3 developed a finance application called Voltron, which saw around 50 banks and companies sign up to test it in 2019, including some of its principal backers. R3 can also boast over 60 ‘partner’ companies which use the Corda blockchain in some way, from small to large.
Corda has also been quick off the blocks when it comes to interoperability. Noting that Central Bank Digital Currencies (CBDCs) are only a matter of time away, Corda announced in June 2021 that it had “developed a solution to enable interoperability between two sovereign networks” which allows cross-network transactions that are as secure as intra-blockchain transactions.
With different countries and banks inevitably running CBDCs on their preferred blockchains, achieving cross-chain communication could prove to be a huge stumbling block to adoption, one that Corda says it has already cleared.
Of course, Corda has not escaped criticism, some of which is fairly fundamental; in 2017 when its code was made open source, reviewers accused it of not being a blockchain at all due to the fact that not all the transactional information was shared between all nodes on the network.
This may be a simple case of semantics, but for blockchain purists it was a further no-no. Corda has also been accused of being too Java-orientated, requiring developers who don’t know Java to learn it in order to develop on the blockchain.
Others have pointed out that with Corda nodes not having a pre-existing transaction chain to work from, processing a transaction that features numerous historical transactions will require the entire history to be downloaded to the node on every occasion of a future transfer.
As time passes and digital assets get new multiple owners, this could prove to be a problem, with transactions needing minutes to process on each occasion.
Corda has also come in for criticism over the way it handles transactions, given that nodes in the Corda network are not set up to verify the uniqueness of a transaction for themselves, which is the case in almost all other blockchains. Corda transaction uniqueness is instead handled by a trusted notary, which could refuse a transaction or accidentally sign conflicting ones, something that blockchains were created to avoid. Regular blockchain validators can, at worst, prevent some transactions from being confirmed.
A malicious Corda notary on the other hand can wreak havoc on the entire ledger by signing conflicting transactions, just one of which would start a chain of incorrect transactions that would send the entire ledger into total meltdown.
And the winner is…
There is a lot to take in when we consider which blockchain is best for enterprises. Much depends on the sector that our notional enterprise is actually in; for example, a logistics firm would benefit from Hyperledger Fabric over Quorum, whereas a bank would opt for Quorum or Corda. We have to bear in mind then that not all three blockchains are suitable for all types of enterprise.
Corda’s speed, scaling and interoperability make it a compelling candidate, while JPMorgan’s Interbank Information Network gives Quorum the edge in terms of a client base compared to its financial competitor. However, Quorum’s comparatively slow speed and the fact it hasn’t made great strides towards interoperability yet mean it has to come in third, despite its illustrious backers.
Seeing as it would be unfair to Corda to use adaptability as a metric against Hyperledger Fabric, we have to go on technical specifications and market reputation. Neither blockchain has concrete real-world performance it can point to, but both blockchains appear to be able to operate in the thousands of transactions per second.
However, Hyperledger Fabric has the edge when it comes to potential – Corda is being developed in-house by R3 developers while Hyperledger Fabric can call on the financial and technical expertise of IBM. If Hyperledger Fabric can get even halfway towards its ultimate goal of 50,000 tps then it could blow Corda out of the water.
Hyperledger Fabric also outperforms Corda in the realm of the present day thanks to its superior market penetration and reputation. IBM Blockchain has become the go-to solution for thousands of companies in all sectors, including financial, and TradeLens has really shaken up the logistics world, and is being used by more and more companies each year.
Corda is undeniably in a better position for adoption than Hyperledger Fabric in its own sector, but until it can boast similar levels of adoption in the finance sector Hyperledger Fabric will remain the number one blockchain for enterprises.