Before we look at Azure, let's look closer at messaging. The importance of timing allows messaging techniques to fall into two main categories: synchronous and asynchronous. With synchronous queues, the messages go in and come out the other end on a regular schedule, like you would on an automated car wash. Cars go in and come out and you don't have any control over how fast the car wash goes once your car is latched in. 

With asynchronous messaging, you can have new messages arriving at random times. They get stored for a while and then are either pulled out of the queue or simply looked at. This is known as peeking. This is more like the grocery store with multiple cashiers and lanes. During a holiday sale, you may have a flood of customers. At those times, the grocer can add more cashiers increasing their ability to handle and store those customers until they can be looked after. Solutions to the queuing problem show up in mathematics or business as queuing theory and in cloud computing too. 

Most cloud vendors have some sort of queue feature to deal with a flood of messages. There is the message passing interface and a host of open source protocols that you could bring in to provide resilience and buffering. 

In Azure, we want to focus on two features, Storage Queues, and Azure Service Bus. That way, you can leverage the built-in functionality to solve messaging problems. We also want to keep in mind the overall subject of long-running tasks.