If anything showed up the fragility of the Ethereum blockchain it was CryptoKitties. The fun, collectible game congested the blockchain so badly that transactions were taking several hours to process and costs skyrocketed for the ‘birthing fee’ of the kitties from 0.002 ETH to 0.015 ETH.
For all the promise of blockchain, one of the most crippling aspects is the speed of transactions. Currently, roughly a maximum of 10 transactions is possible every second across Bitcoin or Ethereum blockchains. When compared with corporate real-life examples of Visa, which typically handles 7,000 in the same period, it is a long way off commercial application.
There are many teams of programmers looking at this problem with Singapore-based Zilliqa an early pioneer.
Headed up by Xinshu Dong, Zilliqa is looking at sharding as a technology to improve the speeds. Dong hails from the National University of Singapore where he did his PhD in computer science, most notably looking at cyber security in complex software systems.
After university, Dong joined Anquan Capital Pte Ltd, which was co-founded by a professor of his, Prateek Saxena, in 2015, together with Max Kantelia and Juzar Motiwalla.
“We originally worked on a bond trading project where technology was antiquated and bonds very illiquid.’ says Dong. “We quickly could see the applicability of blockchain but such applicability was tempered with the slow speeds of existing blockchains.”
The slowness of transactions is caused by the Proof-of-Work (PoW) consensus algorithms by which the blocks are ordered and added into the chain. Every computer or node in the network must agree on whether a transaction is valid or not. This is time consuming.
Sharding tackles the problem by using a ‘divide and concur’ paradigm, explains Dong.
“Not every node has to agree on every single transaction in the sharding model. Say there are 100 nodes, we divide then into ten groups of ten nodes. Each subset is then polled on their agreement and the voting is then aggregated to create a decision.
“However, part of the challenge is deciding how to divide the nodes. If we pick a certain number of nodes to make the decisions then there may be a bias. Or we may open a vulnerability in security if we reduce the number of nodes making decisions.” he says.
Zilliqa uses a PoW process that asks the nodes to randomly elect nodes to make decisions on sharding.
Already they have an alpha version of the product with internal testing reaching a peak of 2,488 transactions per second. To give some perspective on this, if Zilliqa were fully operational on the same number of nodes as Ethereum, the transaction rate would be over 15,000 per second, nearly double that of Visa.
The company recently announced a partnership with Mindshare, one of the largest digital agencies in the world with over 7,000 employees and $31 billion in annual billings. Dong is also talking to other possible customers across a range of industries from payments, to shared economies to airlines.
Zilliqa conducted a private sale and a community contribution in late 2017 and early 2018, raising approximately $20 million. This will be used to run the company for the next three to five years. An open beta version of their sharding technology platform is expected by Q3.
As if that wasn’t enough, Zilliqa is now tackling smart contracts. “This is at least equally ground breaking.” explains Dong. “Smart contracts are revolutionary but there are also many bugs in the smart contract layer – look at the damage that did to the Parity Wallets freezing.
“We are now working on building a secure smart contract language on a fast sharding blockchain. That’s two world firsts.”