A cluster is a replication of the same functional entity. For example, a payment system cluster is composed of 3 servers, and each server runs exactly the same program and has the same function. The purpose of the cluster is to cooperate with the load balancer to share the operating pressure. 
Each node of the distributed architecture is different. For example, the payment system is split into real-time payment services, batch payment services, and reconciliation services. The functions provided by the three services are different from each other, and they cooperate with each other to become a system ecology, and then external Provide services. 
Therefore, the general requirements for the server configuration of each node of the cluster architecture are the same, and there is no communication and collaboration between services. Distributed services often have different server configurations, which are differentiated according to service pressure, and there will be mutual communication and collaboration between services.
Distributed architecture is often combined with cluster architecture to improve availability. For example, the payment system is split into three services. In order to improve the availability of each service, each service must be deployed in a cluster mode. 
The cluster architecture can still operate normally when a node fails, but it is not necessarily distributed.