Process model analysis (
2
)
- Starting model and activation model
First, the starting point model of the process
Anything has the end of the head, a process is no exception, then we now start from the "head" - the process starting point.
First, it is necessary to say that the starting point is also a task node.
Single Start Node
Single starting point is estimated that everyone is more likely to understand that the workflow applications in reality are also most common. As shown in the following figure (1-1), it is a model of single point. In all of the images, a green box represents a starting point.
Picture 1-1)
1.2 starting points
Multi-startflower workflows are not too much in real realization. It is mainly expressed that in the same process, there are a number of starting points. Speaking here, it is necessary to re-disclugly: the starting point is also a task node, but it is not independent of the workflow task. When the process starts, it also needs to complete a specific task to activate the entire workflow. Working.
A multi-starting workflow model basically has three ways.
Multiple solutions
Please refer to Figure (1-2), starting point a and starting point B, which can activate the operation of the process, and after activation, the process will point together to Task A. So, for Task B, it is not concerned about how the process is activated, which is only related to the correct process data from Task A correctly.
Figure (1-2)
Multi-start
In the way (as shown in the following figure (1-3)), the starting point A is activated after the workflow, causing the process to flow along the Task A - Task B - Task C. After the workflow from the starting point B, Task A is skipped.
This way is extremely rare in reality. If you use the process segment of Start B - Task B with Start A - Task B, it is separated from the START A - Task B. The case where the two "subflow" can be approximately viewed (two choices, or multiple choices).
Figure (1-3)
Multiple ways three
Method 3 (as shown below), although there is also a number of starting points, it is basically run in a unified process direction. This is the difference between the two ways. Here, it is necessary to declare this: a start node is also a task node (Task Node). Refer to the Task B / Start B node in the figure.
In this way, in reality, there is still an application. Especially when information is interactive between multiple processes, the flow A sends message data to activate the operation of the process B. However, it is not necessarily activated from the default activation point of the process B, which may be activated from the middle of the process B. For example, the Task B / START B task node in the figure (also starting at the same time)
Figure (1-4)
Second, the activation model of the process
We have seen the starting point model of workflow. Also knowing any processes must have a starting point, or the relative starting point. After a process is activated, it will be carried out from the starting point along the scheduled process route, ordered or disorder (note, I mentioned the word "disorder" here, I will follow the "disorder" state. ).
Therefore, the starting point is the source of this process being activated. Let's take a look at the way a process is activated, or a starting point, activated: from the start of the activation method, there are two ways:
2.1 Artificial activation
Most processes are activated because of artificial information data input or production. For example, a order processing process, the customer submits order information (order information data generation), activates the start of the order processing process.
2.2 Timing or limited time activation
At a specific time, because of a specific condition, a specific condition is contemplated, activating a particular process (or task).
This activation is rarely appearing in reality. Most of the cases, because in a certain process, due to a limited time, there is no need to reach the expected state, and another task or new Process. That is, this method is affected by external factors, and most of them appear together with some process tasks (or process modes).
For example, the order processing flow, the order processing process, limited to 5 days, then defined in the third day, if the delivery notification is not received, an auction information (proactive task) is activated. In this way, the flow system will automatically issue information on the third day.
2.3 External Message Activation
This approach is mostly in the application of multi-process information interaction (or size process nested). Nowadays, the popular business process integration / management (bpm) is basically involved in this area.
As shown in the figure below, the flow A is ended (in all the figures below, the red block diagram, indicating the end node), will send Message to the flow B to activate the operation of the flow B. As for this message is a SOAP message, or by Message forwarded by message middleware, this is different application methods.
In general real-world applications, you need to consider the application interface of JMS or WebService. Most of the workflow applications currently implemented from the individual, mostly using the Message Query mode. Although the development of software, gradually SOA (service-oriented), the security or data correctness of WebService, still needs further development, from this aspect, the height security and message correctness, WebService It is still a little bit a little. In fact, the correctness of security and information is the focus of many application customers very concerned.
However, the development of SOA is a future trend. So now most applications provide JMS and WebService interfaces, or other similar interfaces.
picture 2-1)