Merge "Move "Long Running Business Process" article from Wiki to docs"

doc/source/user/index.rst

@@ -19,6 +19,7 @@ info on concrete features.
     faq
     terminology/index
     main_features
+    use_cases/index
     wf_namespaces
     asynchronous_actions
     wf_lang_v2

doc/source/user/overview.rst

@@ -48,6 +48,9 @@ task should be started at what time. So that Mistral calls back with "Execute
 action X, here is the data". If an application that executes action X dies
 then another instance takes the responsibility to continue the work.
 
+This use case is described in more details at
+:doc:`use_cases/long_running_business_process`
+
 Big Data analysis & reporting
 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
 A data analyst can use Mistral as a tool for data crawling. For example,

doc/source/user/use_cases/index.rst

@@ -0,0 +1,11 @@
+=========
+Use Cases
+=========
+
+This part a collection of articles describing use cases in more details.
+
+
+.. toctree::
+    :maxdepth: 2
+
+    long_running_business_process

doc/source/user/use_cases/long_running_business_process.rst

@@ -0,0 +1,72 @@
+=============================
+Long-running Business Process
+=============================
+
+Introduction
+============
+
+The use case described below is not the most important driver in the current
+development process in Mistral. However, at later stages, when OpenStack
+itself becomes more mature there will be a number of cloud users who would
+want to build enterprise systems following the idea of this use case.
+
+Problem Statement
+=================
+
+Looking back at the industry, say 20 years ago, it’s fairly obvious that
+things have evolved drastically. For example, instead of having one
+information system for everything like accounting, financial planning,
+reporting enterprises tend to have multiple specialized systems in order
+to address performance problems caused by constantly growing amount of
+enterprise data. However, we may need to define a business process, or
+a workflow, that would span several systems and for some calculation steps
+may even require that people interact with the process by entering data
+manually. Those people may be accountants entering primary information,
+office managers, finance directors and others. The formal challenge here
+is to define and maintain a sequence of operations that need to be executed
+one after another and to be able to have some logic (conditions) driving the
+execution of this sequence one way or another.
+
+In order to make this workflow scalable (ability to run some steps in
+parallel), tolerant to failures and observable for external systems
+there has to be some component that would play the role of a coordinator.
+By “observable for external systems” we mean here that we should be able
+to see all the relevant information on how the process has been going on,
+what steps have already been processed and what are left, whether it’s
+stopped with a failure or finished with success. Maintaining a history of
+already finished processes would also bring significant value.
+
+Solution
+========
+
+Mistral is a perfect fit for being this kind of coordinator. In order to
+illustrate everything described so far, let’s consider an imaginary workflow
+of calculating employees’ salaries in an enterprise.
+
+.. image:: img/long_running_business_process.png
+    :alt: Picture 1. Mistral maintains business workflows spanning multiple systems.
+    :align: center
+
+Given an employee full name (or id) such workflow may include the following
+computation steps:
+
+* Calculate salary base using accounting information system.
+* Calculate the employee’s bonus using a different bonus system.
+* Request an approval from a manager for calculated bonus.
+* In case of any error at any stage send SMS to a system administrator
+
+
+In this scenario Mistral always knows the execution state of the entire
+workflow and in case of failures (network losses etc.) it can continue the
+workflow from the point it stopped transparently for a user. Additionally,
+Mistral can run calculation of salary base and bonus in parallel since
+these two tasks are independent. All these things can be flexibly configured
+by a user.
+
+Notes
+=====
+
+The example above is just a simple illustration of what Mistral can offer
+in regard to taking care of long-running business processes. In real life
+Mistral can take care of much more complicated processes spanning multiple
+information systems and involving real people at some points.