A topology is a special type of task. A topology is a collection of tasks that are triggered by events. The execution process of a topology forms a directional acyclic graph (DAG), which starts from the start node and ends at the end node.
A topology can be triggered by a cron expression or an event. Tasks in the topology must meet the following requirements:
The tasks are triggered by events.
The tasks use the Callback communication mode.
The tasks are simple tasks, cluster tasks, or other types of topology tasks.
A task can be added to different topologies and carries different execution contexts in different topologies. Therefore, no cross impact will be generated. If a task is manually triggered or triggered by an event, other tasks in the topology are not affected due to the absence of a map instance.
Elements in a topology
Start and end nodes
Each topology has only one start node and one end node. The task execution process starts from the start node and ends at the end node. Start nodes and end nodes are added in pairs.
Parallel gateways
Parallel gateways are used for concurrent processing and are added in pairs. All branches of a pair of parallel gateways are executed concurrently without conditional judgment. After a branch is executed, the execution process proceeds to the next node after all other branches are executed. For more information, see Use parallel gateways.
Condition gateways
Condition gateways are also called exclusive gateways and XOR gateways. They are used for decision-making in the task execution process, and are added in pairs. A pair of condition gateways have one condition and two branches, which respectively correspond to the true and false judgment results. The branch with the true judgment result is executed. For more information, see Use condition gateways.
Split gateways
Split gateways split task data to dynamically generate data chunks to be executed based on the specified dimension. Split gateways are added in pairs. When you add a pair of split gateways to a topology, you need to specify a splitting dimension for the start split gateway. During execution, the topology can dynamically generate data chunks to be executed based on the configuration obtained by the execution context. For more information, see Use split gateways.
Sample topology
