In most cases, software implementation can iterate fast because developers reuse a lot of existing components. Don't Repeat Yourself—this is a popular rule and motto of many programmers. Using other packages and modules to include them in the codebase is only a part of that culture. What also can be considered under "reused components" are binary libraries, databases, system services, third-party APIs, and so on. Even whole operating systems should be considered as reused.
Backend services of web-based applications are a great example of how complex such applications can be. The simplest software stack usually consists of a few layers (starting from the lowest):
A database or other kind of storage
The application code implemented in Python
An HTTP server such as Apache or NGINX
Of course such stack can be even simpler but it is very unlikely. In fact, big applications are often so complex that it is hard to distinguish single layers. Big applications can use many different databases, be divided into multiple independent processes, and use many other system services for caching, queuing, logging, service discovery, and so on. Sadly, there are no limits for complexity and it seems that code simply follows the second law of thermodynamics.
What really is important is that not all of the software stack elements can be isolated on the level of Python runtime environment. No matter whether it is an HTTP server such as NGINX or RDBMS such as PostgreSQL, they are usually available in different versions on different systems. Making sure that everyone in a development team uses the same versions of every component is very hard without proper tools. It is theoretically possible that all developers in a team working on a single project will be able to get the same versions of services on their development boxes. But all this effort is futile if they do not use the same operating system as in the production environment. And forcing a programmer to work on something else other than his beloved system of choice is impossible for sure.
The problem lies in the fact that portability is still a big challenge. Not all services will work in exactly the same way in production environments as they do on the developer's machines and that is very unlikely to change. Even Python can behave differently on different systems despite how much work is put in to make it cross-platform. Usually, this is well documented and happens only in places that depend directly on system calls, but relying on the programmer's ability to remember a long list of compatibility quirks is quite an error prone strategy.
A popular solution to this problem is by isolating whole systems as application environments. This is usually achieved by leveraging different types of system virtualization tools. Virtualization, of course, reduces performance, but with modern computers that have hardware support for virtualization, the performance loss is usually negligible. On the other hand, a list of possible gains is very long:
The development environment can exactly match the system version and services used in production, which helps in solving compatibility issues
Definitions for system configuration tools such as Puppet, Chef, or Ansible (if used) can be reused for configuration of the development environment
The newly hired team members can easily hop into the project if the creation of such environments is automated
The developers can work directly with low system-level features that may not be available on operating systems they use for work, for example, FUSE (File System in User Space) that is not available in Windows
Vagrant currently seems to be the most popular tool that provides a simple and convenient way to create and manage development environments. It is available for Windows, Mac OS, and a few popular Linux distributions (refer to https://www.vagrantup.com). It does not have any additional dependencies. Vagrant creates new development environments in the form of virtual machines or containers. The exact implementation depends on a choice of virtualization providers. VirtualBox is the default provider and it is bundled with the Vagrant installer but additional providers are available as well. The most notable choices are VMware, Docker, LXC (Linux Containers), and Hyper-V.
The most important configuration is provided to Vagrant in a single file named Vagrantfile. It should be independent for every project. The following are the most important things it provides:
Choice of virtualization provider
Box used as a virtual machine image
Choice of provisioning method
Shared storage between a VM and a VM's host
Ports that need to be forwarded between a VM and its host
Syntax language for the Vagrantfile is Ruby. The example configuration file provides a good template to start the project and has an excellent documentation, so the knowledge of this language is not required. Template configuration can be created using a single command:
vagrant init
This will create a new file named Vagrantfile in the current working directory. The best place to store this file is usually the root of the related project sources. This file is already a valid configuration that will create a new VM using the default provider and base box image. No provisioning is enabled by default. After the addition of Vagrantfile, the new VM is started using:
vagrant up
The initial start can take a few minutes because the actual box must be downloaded from the Web. There is also some initialization process that may take some time depending on the used provider, box, and system performance every time the already existing VM is brought up. Usually, this takes only a couple of seconds. Once the new Vagrant environment is up and running, developers can connect to SSH using this shorthand:
vagrant ssh
This can be done anywhere in the project source tree below the location of Vagrantfile. For developers' convenience, we will look in the directories above for the configuration file and match it with the related VM instance. Then, it establishes the secure shell connection, so the development environment can be interacted with like any ordinary remote machine. The only difference is that the whole project source tree (root defined as a location of Vagrantfile) is available on the VM's filesystem under /vagrant/.
Containers are an alternative to full machine virtualization. It is a lightweight method of virtualization, where the kernel and operating system allow the running of multiple isolated user space instances. OS is shared between containers and host, so it theoretically requires less overhead than in full virtualization. Such a container contains only application code and its system-level dependencies, but from the perspective of processes running inside, it looks like a completely isolated system environment.
Software containers got their popularity mostly thanks to Docker; that is one of the available implementations. Docker allows to describe its container in the form of a simple text document called Dockerfile. Containers from such definitions can be built and stored. It also supports incremental changes, so if new things are added to the container then it does not need to be recreated from scratch.
Different tools such as Docker and Vagrant seem to overlap in features but the main difference between them is the reason why these tools were built. Vagrant, as mentioned earlier, is built primarily as a tool for development. It allows to bootstrap the whole virtual machine with a single command, but does not allow to simply pack it and deploy or release as is. Docker, on the other hand, is built exactly for that—preparing complete containers that can be sent and deployed to production as a whole package. If implemented well, this can greatly improve the process of product deployment. Because of that, using Docker and similar solutions (Rocket, for example) during development makes sense only if it also has to be used in the deployment process on production. Using it only for isolation purposes during development may generate too much overhead and also has a drawback of not being consistent.