A central coordinator node handles requests from all clients.
中心协调者节点会处理来自所有客户端的请求。
Any database partition can be used as or can become a coordinator node.
任何数据库分区都可以用作协调节点，或者被变成协调节点。
Then, the node from which the query originates (the coordinator node) coordinates with the other partition nodes to find all qualifying rows.
然后，发出查询的节点（协调节点）与其他分区节点相协调以找到所有合适的记录行。
If the coordinator node is unable to identify the correct node, it broadcasts the query to all nodes, and then coalesces the responses it receives from them.
如果协调者节点不能识别正确的节点，那么它会将查询广播给所有节点，然后聚集它从这些节点那里收到的响应。
Subsequent requests are treated similarly: the clients always contact the central coordinator node, and the coordinator node retrieves the data from the correct node.
后续请求也被同等对待：客户端总是联系中心协调者节点，而协调者节点会从正确的节点检索数据。
Subsequent requests are treated similarly: the clients always contact the central coordinator node, and the coordinator node retrieves the data from the correct node.
后续请求也被同等对待：客户端总是联系中心协调者节点，而协调者节点会从正确的节点检索数据。
应用推荐