Linear scalability is a subset of the broader concept of scalability.
Specifically, linear scalability refers to the fact that a system can increase its throughput as you provide it with more resources, and the relationship between throughput and resources is linear. In other words, if you were to add 30% more computing power/network/bandwidth to a system, the throughput of the system also increases by 30%. In this way, the internet itself is linearly scalable, as the total amount of network bandwidth available scales directly with the amount of network capacity that is built (to fulfill user demand for more usage).
In DLT systems, those processing resources are nodes. So if a system is linearly scalable, the system’s throughput increases as you add more nodes to the network.
Radix’s Cerberus consensus protocol is the only decentralized Distributed Ledger Technology that achieves practically infinite linear scalability while preserving atomic composability. Radix believes that providing this kind of scalability is the only meaningful goal when considering a public DLT network designed for global finance – any fixed limit on scalability will eventually be reached.