If you have a use case, to process messages of the same KEY in sequence due to the sequence of actions and process messages from different KEYS in parallel, MBQ is for you.

MBQ is an intelligent queue to provide messages based on how you want it and configured your push, Parallel or Sequential.

• Sequential Processing : Provide SequenceKey and all the messages that has same SequenceKey will be processed in sequantial manner based on FIFO order.
• Parallel Processing : Don't provide sequenceKey (or have different SequenceKey) and you are good to consume all messages in parallel based on the FIFO order.
• Queue can be configured to run in a Single JVM with Producer Consumer running as multiple Threads or as a Centralized Service where processing can be distributed across multiple JVM/nodes can connect to the Queue and Consume/Publish messages. Centralized queue is a language independent Queue, Just needs a client to connect to the Queue Server.
• You can schedule messages in the queue. MBQ will provide message to the consumers at scheduled time.
• IN case if your node goes down, your items are safe in a Data store. There are various implementation of MBQ Datastore available to choose from, In Memory, RDBMS, Hazecast, MongoDB etc (hook available to inject your own datastore).

XYZ
(Item #2 - SeqKey A), (Item #1 - SeqKey A)=>Consumer 1
                     (Item #3 - No SeqKey)=>Consumer 2
                     (Item #4 - No SeqKey)=>Consumer 3
                      (Item #5 - SeqKey K)=>Consumer 4
 
 ABC
 (Item #2 - SeqKey A), (Item #1 - SeqKey A)=>Consumer 5

1. You can have Dynamic no of logical queues within MBQ. Queue XYZ and ABC shows 2 logical queues within MBQ.

2. Lets look at Queue XYZ and see how Items are being processed a) Item #1 and #2 are stamped with SequenceKey A -> Only one consumer can consume #1, #2 and in the same order #1 and then #2 b) Item #3 and #4 Doesn't has any Seq Key -> SInce there's no Seq Key, #3, #4 can be consumed by either 1 consumer or multiple consumer at the same time. c) Item #5 has SequenceKey K -> Only one consumer can consume #5 and in the order

3. In case of SequenceKey, Messages will be processed in Order and if any message fails, none of the consumers can pick subsequent message until Failure is resolved or Failed message is moved out of the queue.