2 Managed Code and the CLR

2 Managed Code and the CLR
Remember that managed code is code written for (and with) the .NET Framework. Managed code is called managed because it runs under the constant supervision of an execution engine called the Common Language Runtime or CLR.


Figure 2‑1 Managed Code and the CLR

The CLR is similar to other existing execution engines such as the Java VM or Visual Basic. The CLR supplies memory management, type safety, code verifiability, thread management and security. However, the CLR also bears some important distinctions from previously existing environments.

First and foremost, managed code is never interpreted. This statement is so important that it is worth repeating. Unlike with other execution engines, code that runs under the supervision of the CLR runs natively in the machine language of the host CPU. I will touch on this in more detail shortly.

Secondly, the security model enforced by the CLR on managed code is different than the security of previous environments. Managed code does not run in a virtual machine, and it does not run in a sand-box. However, the Common Language Runtime does apply restrictions to managed code based on where the code comes from. Managed security is flexible and powerful feature. An entire tutorial in this series is devoted to this topic.

Before beginning to describe what and how the CLR does what it does, I would like to take a moment to address the need for managed code. Managed code makes it possible for applications to be composed of parts that they were never tested with, and that may not even have existed at the time of development.

For example, Acme Widgets sells their product with a website online. WeShipit Delivery Service provides shipping. If Acme publishes a shipping interface to their website, then WeShipit could become a shipping agent for Acme’s site. It is feasible, with managed code, for WeShipit’s shipping code and web interface code to plug right into the website for Acme without anybody at Acme lifting a finger, or even being aware of the addition to their site.

This kind of flexibility will create some exciting opportunities for the composition of software across the Internet. (Imagine an online game where the players can write software to define their characters and their characters’ items). However, it raises concerns of practical details like faulty or malicious components, performance, type compatibility and type safety. These are the reasons that a managed environment is necessary. However, the last thing software developers want is to suffer the performance-hit of an interpreted environment.