Concurrency – An Introduction

–What are concurrency and parallelism? Why should we study them? In this chapter, we will look at the many aspects of the concurrent programming world. The chapter starts with a brief overview of parallel programming and why we need it.  We cover ground pretty fast here, touching upon the basic concepts. 

As two major forces, huge data size and fault tolerance drive concurrent programming. As we go through this chapter, some of our examples will touch upon some clustered computing patterns, such as MapReduce. Application scaling is an extremely important concept for today's developer. We will look at how concurrency helps applications scale. Horizontal scaling (https://stackoverflow.com/questions/11707879/difference-between-scaling-horizontally-and-vertically-for-databases) is the magic behind today's massively parallel software systems. 

Concurrency entails communication between the concurrent entities. We will look at two primary concurrency models: message passing and shared memory. We will describe the message passing model using a UNIX shell pipeline. We will then describe the shared memory model and show how explicit synchronization creates so many problems.   

A design pattern is a solution to a design problem in context. Knowledge of the catalog of patterns helps us to come up with a good design for specific problems. This book explains the common concurrency design pattern.   

We will wrap up the chapter by looking at some alternative ways of achieving concurrency, namely the actor paradigm and software transactional memory.  

In this chapter, we will cover the following topics:

• Concurrency
• Message passing model  
• Shared memory and shared state model
• Of patterns and paradigms