The Definitive Guide to Containers

Containers are self-contained, isolated units of an application that share the kernel of an operating system but do not contain their own operating systems. This shared kernel acts as its core module, keeping it very lightweight. This means each container can be distributed across a host environment without needing a unique virtual machine for each container. 

What containers do contain is all of their own components needed to run anywhere, including libraries, dependencies, and code. These are all the nuts and bolts they need to function independently, on any platform.

Containers are at the heart of IT modernisation, microservices, and infrastructure-as-a-service (IaaS) adoption. That’s because they’re in lockstep with the mobility, flexibility, and efficiency these movements embody. Common benefits of containers include:

• Affordability. Containers can be a more cost-effective way to distribute workloads across computing resources. By increasing deployment density, you’re making the most of existing resources. Also, minimum costs to run containers can average less than a hundred dollars per container while the cost to run virtual machines can range more in the hundreds per VM.
• Performance. Containers can start up faster than an application on a virtual machine, which can sometimes be seconds rather than minutes. 
• Security. By design, containers are self-contained, which means vulnerabilities within them may be more contained as well. Whereas a monolithic application provides a larger attack surface area, containers are many small ones—and this can help cordon off the effects of a vulnerability.
• Efficiency. Containers use far fewer resources because they don’t require their own server or virtual machine. They’re smaller, so you can run more apps concurrently on a single server. The shared kernel means they’ll also require less memory to start up than an application that needs to boot up an entire operating system
• Modularity. Single components can be added, updated, or patched without having to redeploy a monolithic, tiered application. This reduces disruption to the user when updates are made, as well as the likelihood that updates to a single component could break or affect other components. And, distributed apps can be more failure-tolerant compared with tiered applications.

Containers compartmentalize and improve application development, which makes them beneficial to organisations’ infrastructures, the teams operating them, and the end users of applications and services.

• On the back end: Containers can improve how applications are organized, stored, and updated, and with leaner operating costs. Containers also revolutionise how IT provisions servers and IT resources, occupying a much smaller footprint with self-sufficiency that stands up to even the most agile environments. In terms of DevOps’ needs, containers can get IT teams off the hook for having to provide many of the quick-turn resources needed to move in fast development cycles.
• For developers: Containers boost efficiency, speed, and agility, all of which lead to faster development cycles, fewer bottlenecks, and more productivity. As teams leverage the cloud more, containers are practically designed to thrive in the cloud with agility and mobility by design. Also, containers remove some of DevOps’ dependencies on IT so they can focus on applications and on-the-fly development.
• For end users: Better application performance and seamless updates and patches that can happen without full-scale reboots mean a better user experience.

Containers were designed to be stateless—an inherent characteristic that made them perfect for some scenarios but tricky for others. Stateless means impermanent—and not storing data for longer than the life cycle of the container. This meant they were best suited for disposable, short-lived use cases.

However, being able to store and access data is what makes most applications valuable. Developers long struggled to leverage the benefits of containers for workloads that clashed with their design principles (or design limitations).

Today, stateful workloads are the norm for containers. And storage is a very important component of a successful container management strategy. This is thanks to container-native storage engines and platforms like Portworx® that have evolved to support stateful workloads. Now, containers can even be used for traditional, legacy applications.

Container management is the processes, software, and tools used to create, deploy, monitor, and manage containers along their life cycle—much of which is abstracted away and handled by container orchestration platforms such as Kubernetes. Within container management is container orchestration, which includes the process and tools used to manage a fleet of containers, including how they interoperate and communicate with one another on a broader scale.

Container orchestration has become largely automated to streamline the work required to spin them up and down, which can be considerable when they begin to multiply in an organisation. Container orchestration can also provide some standardization, security, and governance that benefit large enterprise organisations implementing container strategies.

Docker is a runtime environment used to create and build software inside containers. As we mentioned above, containers include all the individual components they need to run anywhere—known as a container “image.” Docker is a platform that enables you to package up those components.

The key to Docker’s simplicity—and what makes it so popular—is how it enables declarative programming. Meaning, as long as you know what the outcome of your container should be, Docker can help you achieve and optimise that outcome.

Docker also offers built-in security services and encryption measures, which can help address some primary concerns with container security.

Kubernetes is one of the most popular container orchestration systems. What makes it so critical to deploying containerized apps at scale? Automation.

When you have thousands of containers or more, managing container sprawl and complexity is much more manageable with a container management platform. Kubernetes streamlines and automates aspects of container orchestration so DevOps teams can seamlessly orchestrate complex container-based architectures. It can even make decisions for you about where workloads should run and move them if they’d run better elsewhere.

Benefits of Kubernetes include:

• Presets that keep containers running against “desired state management” and other criteria. Deviations from defined criteria are autocorrected.
• Scaling on autopilot
• Monitoring, health checks, and other automated maintenance operations that ensure better availability
• Autorecovery and rolling restarts. That means less time spent fixing issues such as bottlenecks, failures, and outages.
• Automation of tasks such as load balancing, restarts, scheduling, and scaling

Learn more about Kubernetes

Portworx by Pure Storage® is a data services platform for Kubernetes that helps you manage all things related to container data. Portworx provides a fully integrated solution that enables you to deliver persistent storage to containerized apps. It also handles data protection, disaster recovery, data security, and automated capacity management for applications running on Kubernetes. Essentially, it takes care of the concerns of container management for you.

Learn more about Portworx

Virtualisation was then—containerization is now. However, practically speaking, when you’re running applications in the cloud, containers and virtual machines together aren’t always an “either/or” scenario. They can accomplish very different things, with different benefits, but they’re often working in tandem. These are complementary technologies that will both come into play as you deploy apps in the cloud.

The key difference between the two technologies is their goal. VMs require their own operating systems while containers share the same OS kernel. VMs are all about consolidation in the data centre, preserving more servers and IT resources. Containers are more about the nuts and bolts of application management—the day-to-day of DevOps. And, more often than not, containers are doing what they do with the help of a VM or hypervisor.

Learn more about containers and virtual machines and the tradeoffs between them

Portworx by Pure Storage is a Kubernetes data services platform built from the ground up for the modern era of containerized workloads and microservices. Portworx provides a data services platform for cloud-native applications running across hybrid-cloud satellite locations. The platform offers solutions designed to address scalable container storage, backup and disaster recovery, multi-cloud operations, data security, and capacity management. Portworx also helps clients address their internal policy compliance and governance.