xfsm

Library goals

• distributed competing consumer finite state machine
• integration with the major relational database (sql server, oracle, postgres)
• if efficient, an integration with a distributed memory cache like redis
• porting on at least 2 main programming languages (c#, java)
• excellent error handling
• states transitions will be defined by the business logic (at runtime) during the item elaboration
• provide rest APIs for stats (ie: tot. number of elements, amount of elements by status, last errors, ecc...)
• provide a rest api interface which provides the same statistics as above and add some interationcs features (ie. add new elements, retrieve a specific element, etc.)

Restrictions

• Collection items must be of the same complex type

What the developer needs to do in order to adopt this lib

• states definitions: has to implements the required IXfsmState concrete business logic
• states transitions: into the business logic, when required, has to use the lib api in order to move the item in the next state

Library "must have"

• fsm: generic finite state machine library processor features; will "move" the item state from A to B (with proper error handling)
• consistency and durability: every state transtions will be stored in a relational database
• concurrency: the library will garantee thread safety among elaboration sessions
• competing consumer: in a distributed architecture with multi process/thread condition, will guarantee that a specific item X will be processed by a single consumer at a time
• data rolling: after a while (defined by configuration) items in a "final" state will be removed from the pile

Use cases

• Generic sales report (xml document elaboration):
  • Subject: an xml document containing the employee selling results for the day before
  • Steps to execute:
    1. validate the document in order to catch syntax/semantic errors and also some business domain validations (ie. the user exists, etc..)
    2. save in database some document information (user, stock sold, amound, currency, ...)
    3. generate a pdf report available for the management
    4. delete the document
• Industry 4.0 generic sensor event:
  • Subject: a sensor raised an event, the informations sent to our application must be managed
  • Steps to execute:
    1. analize data arrived from the sensor and take the right action (turn on the alarm, send an sms, etc...)
    2. archive event data for auditing purpose
    3. update some statistics based on event type

When this lib can be of any help?

• When you don't have any chance to rely on some other well known products/technlogies (ie. RabbitMq, Kafka, ecc..)
• When you simply can't move your business logic on any other technlogy rather than your RDBMS.
• When you simply don't want to introduce new elements in your architecture to undestand/manage/monitor/etc...
• When you basically have a strong know how on your RDBM in your team and you're confindend on this for the moment.
• Maybe you just undesrstood there's the necessity to decouple your services/business logic and moving over using this lib can be a little step required than introducing some other components

Disclaimer

This is not intended to be a sobstitute of any other queing or competing consumers technlogies. Is just a case study hoping that can be of any help someday, somehow to someone :)