The debate over blockchain scalability has been raging for years, and has created some of the most entrenched positions – and forks – in the cryptocurrency community.
From the Bitcoin Cash fork of Bitcoin to the controversy surrounding the initiation of SegWit2X, which eventually became SegWit, scaling has taken front and center in Bitcoin.
Beyond Bitcoin, however, scaling has extended to other blockchain networks as well – such as Ethereum.
Looking past the heated debate among the communities on how to scale public blockchains, reveals a two-sided camp: on-chain vs. off-chain scaling.
On the layered, off-chain side is Bitcoin’s path with its young Lightning Network. On the other, on-chain side, is Ethereum, with its trajectory towards Ethereum 2.0, also known as Serenity.
Zooming in on Ethereum, Serenity’s ultimate goal is to become a proof-of-stake (PoS), sharded public blockchain.
“Next-generation” blockchains, such as the recently launched Cosmos, have focused on shifting the consensus design of Bitcoin (proof-of-work) to one that is more scalable and energy-efficient with PoS.
However, Ethereum plans on combining PoS consensus with another emerging technical development – sharding.
Sharding is not explicitly a blockchain invention. The concept has been around, and implemented continually, as a method for horizontally partitioning databases.
The data is stored on individual “shards,” which are the partitions that store separate pieces of a database – sometimes different shards will all store the same bit of data, such as an identifier.
Some advantages are derived from sharding databases, ranging from the performance of networks to storage savings.
As blockchains are distributed ledgers of data, sharding might make sense for improving their performance. In particular, for public blockchains as opposed to enterprise blockchains.
The primary difference between public and enterprise blockchains, is critical to understanding the role of sharding in blockchains and cryptocurrencies.
Public blockchains are “permissionless,” meaning that anyone can join the network and run a node to participate in consensus – they can even do so anonymously in some cases.
Conversely, enterprise blockchains are “permissioned,” which is where node operators and consensus participants are approved or publicly known by a consortium of entities.
For example, Hyperledger Fabric, a popular open-source blockchain framework, has a customizable, permissioned voting consensus engine based on variations of Byzantine Fault Tolerance.
As a result, a consortium of say, investment banks, could operate an enterprise blockchain using Hyperledger Fabric, where they act as the gatekeepers and ultimate arbiters of what is appended to the blockchain.
Enterprises (i.e., permissioned) blockchains can have a much higher transaction throughput, because their consensus is faster and less susceptible to malicious attacks – since all parties are known and working towards similar goals.
Enterprise blockchains are undoubtedly a popular option among businesses, but Gartner recently revealed that 90 percent of these implementations will need replacement by 2021.
Why? The value proposition is decidedly different for a public blockchain, which requires a much more nuanced, and advanced, engineering approach. One that will eventually leave enterprise systems behind.
Permissionless (i.e., public) blockchains need to find ways to scale to higher transaction throughputs, without sacrificing their decentralization or ability for users to enter and leave the network at will.
Ethereum may be the public blockchain carrying the torch into sharding, but an entire class of new blockchain platforms has emerged as competition.
Sharding blockchains is an area still under heavy active research and is an esoteric field at the edge of distributed computing. Ethereum researchers and other blockchain projects are pursuing its potential rewards, nonetheless.
“Despite lofty promises, creating a decentralized, secure and scalable public blockchain has proved to be a strenuous task,” says Beniamin Mincu, CEO of Elrond.
“Sharding is complicated, but it provides the type of throughput capacity improvement that enables public blockchains to rival networks like VISA. Some of its challenges, though, include single-shard takeovers, cross-shard communication and data validity.”
Elrond has a live testnet for a model known as ‘state sharding’, combined with a secure PoS consensus.
Also, Singapore-based Zilliqa has recently went live as a sharded blockchain, claiming it will support their vision for a better, safe and easy smart contract network.
The challenges that sharding face are some of the reasons why many Bitcoin proponents opted for second layer scaling in contrast to protocol level – it adds layers of complexity.
Though, that has not stopped researchers and public blockchains from working towards a realization of sharding, which is more attuned to the performance requirements of a smart contracts platform like Ethereum.