Shardeum, a L1 blockchain network, is gearing up for an important milestone on 5th November 2022 just weeks after announcing the completion of its seed funding round that saw it raise $18.2mm. The network will demo a key feature from its innovative technology — auto-scaling. This blog will explain why auto-scalability is crucial to Shardeum’s goal of overcoming blockchain trilemma which in turn will help it to keep the gas fees very low and constant on the network.
Milestones Achieved So Far
Shardeum was launched in February, 2022. The project’s protocol layer is called Shardus which is developed by Omar Syed and his team since 2017. Shardus previously had demonstrated auto-scaling and linear scaling features which were key elements to launch Shardeum and its mission to onboard billions of users to Web3. This was possible with the help of sharding and more specifically, dynamic state sharding which is considered an advanced version of static sharding used by more recent blockchain networks such as Near and Elrond. The protocol further completed EVM integration earlier in 2022 making Shardeum an EVM-based layer 1 smart contract platform that can scale linearly with atomic and cross-shard composability.
After Shardeum Foundation was structured in Q1 of 2022, the project set its sights on successfully demonstrating linear scalability to its community for the first time as a layer 1 open source blockchain platform. Linear scaling, in simpler words, means the capability to scale up by adding more nodes to the network. And with more nodes, project aims to not just scale infinitely but expand horizontally by allowing average users like you and me to run a node on the network and keep it safe in return for rewards. April 2022 marked the first time linear scalability was demonstrated by a blockchain network. The below images were taken from the demo event that shows the TPS with 100 nodes and the TPS with 200 nodes. Apologies for the poor quality of the images.
Understanding How Transactions are Processed on Shardeum
Let’s start this interesting exercise briefly by understanding how transactions are processed on Shardeum. Unlike typical blockchain networks, nodes in every shard on Shardeum will validate each transaction individually and reach consensus on them. With dynamic state sharding, nodes are assigned multiple address ranges so when consensus happens on individual transactions, network will process them simultaneously. Such transactions are then batched together in partitions and passed onto archive nodes on the network who are responsible to maintain the transaction history. In typical blockchains, transactions are first added to a block and consensus is applied at a block level which means processing can only happen sequentially slowing down the network. Moreover, the recent blockchains with static sharding run into the problem of waiting to form a new fixed shard before processing every transaction.
Validator node on Shardeum will thus need to download only the current state of the network pertinent to the transactions they handle making it affordable for average users to run nodes and keep it safe. Light architecture as such facilitate Shardeum to take advantage of its other breakthrough features including auto-scaling. Note, although auto-scaling is not new, incorporating this feature at the protocol level will be the first time ever by a layer 1 blockchain.
Shardeum & Auto-Scaling
Asides from its capability to scale linearly, Shardeum will make use of one of the hallmarks in its technology — auto scaling — which will allow the network to automatically increase or decrease its TPS (throughput) capacity based on demand in the network.
Autoscaling will work by measuring the network load every cycle (60 seconds) and coming to consensus on the required number of validator nodes needed to process the current load. This is similar to the Bitcoin network coming to consensus on its difficulty level. For example, when an application on Shardeum goes viral, the network will react by independently adding more active validator nodes to increase throughput capacity. If traffic on the network goes down, it would shrink the number of active validators and keep the operations on the network sufficiently nimble. Remember, auto-scaling is self executed without any need for intervention from an intermediary.
Why Auto-scaling Matters?
Auto-scaling is important for quite a few notable reasons. Firstly, when you build a network, it should ideally be able to self-govern the number of nodes with an optimal and dynamic incentive mechanism. Maintaining high efficiency while scaling to meet demand is what will help keep the cost of the network and ultimately the average transaction fees low. Secondly, auto-scaling helps to maximize decentralization and security. Shardeum, as part of its design, will randomly rotate the validator nodes in and out of the system. When a validator node is rotated out, a ‘standby node’ is rotated into the network to validate transactions. And when the traffic is on the move, the network will make use of ‘standby node’ concept accordingly. And finally, as previously indicated, auto-scalability and load balancing are techniques leveraged by traditional technology powerhouses to seamlessly scale up their networks as more users joined them.
Sharding is the Solution
Blockchain was introduced to bring maximum decentralization and security in our daily lives after the 2008 financial crisis and the focus was never on scalability. But with the technology thriving despite all the challenges it faced through the last decade, industry experts like Vitalik Buterin rightly recognized the need to scale up the Web3 ecosystem to not just compete with Web2 peers but also to replace them in the near future. That’s when solutions like sharding and parallel processing were theorized as potential answers to scalability trilemma through which blockchains can become more efficient. When compute workload, storage, and bandwidth are evenly distributed across all nodes on the network, it will naturally save more resources. That, in turn, will allow the networks to consider and leverage other scalability solutions like auto-scaling for instance.
No comments:
Post a Comment