Automation of Node Redundancy for Stable Operation of Various Types of Blockchain Nodes

This paper describes automation of node redundancy for stable operation of many types of blockchain nodes. In recent years, traditional large companies have gradually begun to consider the use of public blockchains. When using blockchain for mission-critical purposes, it is necessary to operate a large number of nodes to increase redundancy and ensure stable operation. However, as public blockchain technology is rapidly advancing and the number of types is increasing, it is difficult for many companies to efficiently operate nodes. To overcome this difficulty, Ginco Inc. developed a system on the cloud that mainly consists of three main technologies. By automatically creating a snapshot of a volume containing chain data and attaching it to a new node, Ginco Inc. reduced the time required to start up a node. Additionally, in order to prevent requests from clients from being routed to nodes with low chain block heights, Ginco Inc. implemented a function for the load balancer to grasp the block height of each node. Furthermore, when reassigning clients to nodes due to a node failure, Ginco Inc. is able to do the reassignment evenly by using consistent hashing as the mapping algorithm. Using these automatic redundancy technologies, Ginco Inc. was able to implement node services for about 30 types of blockchains with just a few infrastructure engineers. The node services have continued to operate without any large system failure for over 5 years since its launch.