Dynos are intended to be easily instantiated, configured, replicated, and non-persistent, meaning that they can be easily spawned or destroyed without consequence. Dynos exist to serve a particular service rather than act as long-term data stores. In this way, Dynos are analogous to dynamic application servers in traditional cloud architectures, that can scale up or down depending on network load. The main difference here being that any Node in the Zero Grid could choose to provide these services, rather than a centralized company. This enables the Zero Grid to use the speed, availability and scale of existing cloud providers for certain use-cases, without sacrificing decentralization.