When it comes to facilitating transactions, advocates of Blockchain technology often point to the fact that transactions come with a built-in, fail-proof layer of security, thanks to the inherent nature of the ledger on which the transactions are recorded.
Because of this, Blockchain is being heralded as the future of secure and efficient transactions, as it could eliminate the need of trust on any central party for validation and bookkeeping.
While Blockchain certainly brings a paradigm shift to many businesses and applications, there’s just one “minor” problem: scalability.
This problem is nothing new. A quick Google search on the topic of Blockchain scalability produces hundreds of results from forums, websites, and blogs that all describe the same problem. While cryptocurrency is enjoying more and more mainstream adoption, it does not process transactions fast enough for it to be used on a massive, or even near massive, scale. As recently as June, Bitcoin, one of the most prominent cryptocurrencies so far, could only process a maximum of seven transactions per second (tps) under optimal conditions, with more realistic numbers looking like two to three, compared to PayPal’s 115 and VISA’s 2,000 tps.
As revolutionary as it has proven to be so far, if Blockchain technology is to re-shape our financial system at its core, it must be capable of scaling out and conducting at least as many transactions per second as networks like VISA do today.
There are a number of promising projects that could help Ethereum scale to match the transaction rate of VISA, including Plasma and Raiden, but these projects are providing alternative off-chain or side-chain channels, rather than addressing the scalability of the main Blockchain itself.
Looking at other major Blockchain platforms (who have demonstrated actual results), transaction speed depends upon protocols and functionality.
IOTA can apparently now handle 500-800 tps, Waves has hit a few hundred transactions per second and Ripple has reached over 13 tps.
NEO is using a technology called Delegated Byzantine Fault Tolerance and can currently reach 1000 tps.
The Blockchain platform Zilliqa demonstrated the capacity of processing 2488 tps with an approach it calls “sharding,” a clean-slate protocol that is built to scale in an open, permission-less distributed network that does not compromise resilience and security.
What is sharding?
In order to perform a transaction on Blockchain today, all computers, or nodes, in the network must validate the transaction or execute a smart contract, a piece of code stored in the network that delineates the conditions necessary for the transaction to be carried out. If all nodes achieve the same result and reach a consensus, the transaction is confirmed. As you can imagine, this takes time.
Sharding, on the other hand, runs on the parallel processing power of multiple networked machines that split up the workload of verifying transactions. It automatically divides networks into smaller sections, or “shards,” each of which runs a smaller-scale consensus protocol.
Processing in parallel, such a network is capable of churning out hundreds of transactions per second per shard, for a total of thousands of transactions per second.
As more nodes join, the network will become increasingly faster at validating transactions. Once such a network is as large as that of Ethereum, it will be able to handle so many transactions per second that it makes it faster and potentially cheaper than VISA.
If such systems achieve the speed necessary to take on the workload of our current banking system without sacrificing the permission-less nature of distributed networks, dApps that make use of everything from auctions to payments will be able to operate on its robust, secure, and efficient protocol. Sharding may help facilitate a major breakthrough in solving Blockchain’s scalability crisis.