The original implementation of the network node – launched in September 2017 – has taken us as far as it could. We’ve known for a long time that a new architecture is needed to achieve our roadmap, prepare the system for Shelley, and provide a foundation for Goguen, as well as other future developments.
cardano-node design is the first to take advantage of our work on formal methods. While the old node was monolithic – with components like the wallet backend and explorer built in – the new version is modular. This makes future integrations easier and allows the node to be more readily incorporated into other systems, such as those used by exchanges. In the new architecture, the node, wallet, and explorer exist as separate components (a new wallet backend will soon be released).
A significant achievement of this new implementation is the separation of the consensus layer and ledger rules. This decoupling means we are able to change the ledger rules without making changes to (or risk breaking) consensus. Following from this, when we transition into Shelley to Goguen, only the ledger rules will change. This will allow us to execute deployments more efficiently and add new features more frequently. We’ll have less to validate and test, while supporting more efficient development.
Some benefits will be immediate, and others will be realized over time. The direct benefits are that IOHK engineers will be able to innovate more easily and make changes to specific components without necessarily impacting others. The new implementation, coupled with the update to Ouroboros BFT, will also lead to significant TPS (transactions per second) performance improvements. For end-users, the benefits of this update will be cumulative, as the Cardano network profits from greater developmental support and system adaptability and portability.
This new implementation is the result of a lot of hard work. Now, we start to see the benefits of our commitment to formal methods, delivering a network that can not only scale, but remain stable while doing so. The new codebase has had substantial – and ongoing – testing, and we’ve been able to make a number of fundamental improvements without inheriting the shortcomings of the old codebase.
The new Cardano node also features an IPC interface that can be used by multiple client components, including wallets, explorers, CLI tools, and custom integration APIs and tools. This isn’t only about us being able to develop better-performing systems and applications, but others being able to as well.
Note: The new cardano-node is listed on Github.