Welcome to the appLariat blog! We'll be focused on discussing technical and business topics related to application containers and hybrid cloud, but I'm sure you'll also see some interesting thoughts on the tech industry and business in general.
As we are preparing to launch the company and product at DockerCon 2017 in Austin, I was reflecting on how the company got started and for our first blog, I thought I'd share why we started appLariat and built our Container Automation Platform.
When my co-founders and I first started brainstorming ideas for our next company, application containers came to the forefront quickly. All of us had experience with first generation Cloud Management Platforms (CMPs) that could spin up virtual machines in multiple clouds and had seen the issues inherent in dealing with automating deployment of applications. All the CMPs that tried to address deployment and management of applications required a "blueprinting" process to define an application to the system. It was a time consuming, laborious process and blueprinting had a lot of problems. For one thing, application blueprints were not portable and had to be modified to work on different cloud environments. Another issue was that the way a deployment was done for each component of an application:
- Fire up a virtual machine in the chosen cloud
- ssh into the virtual machine and manually install a middleware component of the application stack
- Upload a code, config or data artifact.
In real world implementations of these blueprints, manual component installations would fail and you'd have a bunch of orphaned virtual machines with no working application. Or you'd spend considerable time writing orchestration scripts that could detect failures and clean up orphaned VMs and notify the user. For this reason, most first generation CMPs were just used to fire up virtual machines in the clouds they supported. Installation of applications was either done manually, or through tons of Puppet, Chef or Ansible scripting by the ever-growing DevOps team.
We believed that application container technology could solve a lot of the issues related to application blueprinting, including providing cloud portability. We also saw that container orchestration runtimes like Kubernetes were maturing and could provide scalability and resilience for applications. It was clear that the next generation of CMPs should be based on containers running in a container orchestrator.
So, we started doing some research and found a few things. Container-based application deployment solutions, some with built in CI/CD, were in the market. Some had built their own container orchestrator (most of those companies are gone today) and some used Kubernetes, Swarm or Mesos. But quickly a few things became clear:
- To deploy applications using these solutions, you needed to provide Dockerfiles and Docker Compose files. This meant that the developers had to have already built the application to run in containers. That meant that they'd already dealt with the steep learning curve for developing applications in containers and it also meant that the solutions were best suited to new applications built cloud native.
- While Kubernetes and the other container orchestrators provide a lot of capability, knowing WHAT to do, they don't often know WHEN to take action. Scaling clusters up/down to meet the needs of deployed applications, as well as lease termination for different workload types are a few examples of things that the orchestrators need to be told to do.
- Kubernetes was the clear leader in container orchestration.
Now, while containers are all the rage in Silicon Valley and the largest financials, and they can afford the time and talent required to master container technology, we started thinking about mid-market companies across the country (and the world) that run their business on existing applications that are in ongoing development. In our research, we found that many of these companies had limited exposure to containers and container orchestrators. They did not have the resources to get up to speed on these technologies, but when we explained the possibilities of what containers could do for them, they were very interested. There also many large enterprises that do have familiarity with container technology but don't know how to scale it for hundreds, even thousands of applications.
What was needed was a Container Automation Platform, that could allow these companies to get cloud native capabilities for their existing applications through the power of containers, without the steep learning curve for technologies like Docker and Kubernetes. The idea for our new company was born! We were going to democratize container technology for any company that runs it's business on applications they develop.
We decided to focus on two differentiators:
- Automate containerization of existing applications with no need for the customer to learn Docker or manage Dockerfiles or Docker Compose files. Our goal was to allow someone to use our product to create a container model of an existing application in minutes that could be deployed on Kubernetes in a public cloud or a customer's existing vSphere environment
- A policy engine to control the WHEN for Kubernetes in support of deployed applications
Now, of course, we provide a complete turnkey container automation solution that can silently install the Kubernetes orchestrator, automatically deploy applications and integrate with a company's existing development and release processes and tools. We also provide application and cluster capacity monitoring, presenting the data in measures that IT Operations teams are familiar with, rather than the more esoteric data directly available from Kubernetes. But the key to the appLariat Platform is that you don't have to know anything about containers to get your application running in containers on Kubernetes. We handle all the complexity for you.
So that is how our company came about. We believe that democratizing container technology is our mission and we couldn't be more excited about launching our company and product at DockerCon. If you're at the show, come by and see us!