Decentralized Consensus Algorithm For Scalable Blockchains

Abstract

Blockchain is a decentralized distributed ledger. Today blockchain technology is one of the most researched and adapted to modern systems due to its security advantage. The digital currency Bitcoin is the living proof of how much blockchain technology can provide reliability and trust into any system since it is not managed or governed by any organization or person. But when looking into such systems which are decentralized, there are shortcomings like scalability. This research proposes a consensus algorithm for private blockchains that can scale with increasing number of ordering nodes. Currently almost all blockchains have a threshold limit of how much transactions that can be processed in a given amount of time. Blockchains that are public and include massive amount of users in the network process transactions at a very slow rate. Bitcoin and Ethereum process 7 and 15 transactions per second respectively. VISA processes around 25000 transactions per second. Therefore, blockchains are still not at a level that can replace the traditional financial systems. This case is similar to that of private blockchains. Even though they are faster than public blockchains, then again the requirement for in systems that limited in the same manner are even greater than what current blockchain implementations can provide. The Canopus consensus algorithm proposed here promises scalability to increasing amount of ordering nodes in the network. It is tested on the Hyperledger Fabric framework version 1.4. A sample smart contract is instantiated for all the tests. The performance was evaluated for different amounts of ordering nodes.