Ten Docker projects that are rapidly evolving

Docker is undoubtedly the most popular open-source technology this year, and Docker is now the darling of entrepreneurs and innovators in the IT world. Regardless of Google, Microsoft, Amazon, IBM and other technology manufacturers are actively supporting Docker technology, Docker although the introduction and use is very simple, but the whole ecosystem is quite large, and its underlying technology is also very complex, the current based on Docker technology projects springing up, today, The author summarizes the rapid evolution of the Docker related technology, to share with you.

Kubernetes

In the summer Dockercon, Google Infrastructure Vice President Eric Brewer announced Kubernetes, built on Docker Kubernetes can build a container scheduling service, The goal is to enable users to manage cloud container clusters through the Kubernetes cluster without requiring users to perform complex setup tasks. The system will automatically select the appropriate work node to perform the specific container cluster scheduling processing work. Its core concept is container Pod (container silo).

Docker Pods

Closely related to Kubernetes, Eric Brewer also talked about containers and introduced the concept of "pods". This is a very key concept in kubernetes. He said: "At Google, we rarely deploy a container, instead, they combine the containers together." ”

He pointed out that the Docker container needed a constant mapping of internal and external ports, a complex additional layer that needed to be managed, stored, and queried-even between pod containers deployed as a single unit. So, at Google, they make sure that each container's pod has its own IP address. This means that the port used is the same both internally and externally. This removes the complex add-on layer of the management port. Now, find pods running a specific service and you only need these pods IP address lists. The Google Computing engine is currently the only cloud infrastructure service that helps allocate an IP subnet to virtual machines-so there is only one IP per docker pod.

Weave

Weave is developed by Zett.io, which can create a virtual network to connect Docker containers deployed on multiple hosts. By weave all containers like being connected to the same network switch, applications that use the network do not have to configure port mappings and links. External devices can access the services provided by the application container on the Weave network, while existing internal systems can also be exposed to the application container. Weave can penetrate the firewall and run on a partially connected network. In addition, weave communications support encryption, so users can connect to the host from an untrusted network.

CoreOS

CoreOS is a new, architecture-redesigned Linux distribution that can run on existing hardware or cloud. CoreOS uses SYSTEMD and fleet to manage the container, and to share information through ETCD service discovery and configuration. CoreOS is currently in the limelight, has been financing and announced the acquisition of Private Docker warehouse service Quay.io, into the enterprise-level registry. In addition, CoreOS's ETCD and other components have also been recognized by the community and have been used on a large scale. CoreOS has released its first stable version, and the current mainstream cloud service providers provide support for CoreOS.

Flannel

It is an overlay network tool designed by the CoreOS team for Kubernetes to help each CoreOS host that uses Kuberentes to have a complete subnet. The name before flannel was that Rudder,kubernetes would allocate a separate IP address for each pod, so that the containers of the same pod could be connected to each other, and the previous CoreOS had no such ability. To address this problem, flannel sets up a subnet for the host by creating an overlay network in the cluster.

Docker for Windows

Recently, Microsoft has joined the Docker, saying they intend to build a containerization solution for Windows and provide a Docker-compliant API on top. Although Docker images can never be portable between Linux and Windows containers, it does mean that there are already some tools available to build on the Docker API layer.

Atomic

Project Atomic is a recently released item that is also a thin version of Linux that contains only Systemd/geard/rpm-ostree and Docker components that are designed to deploy and manage Docker containers. It can run a large number of containers at a high performance near the hardware bare-metal level, and it is also based on SELinux and secure.

Cloud Foundry Diego

The focus of ActiveState is the open source PAAs project. We believe that the Docker integration caused by the Diego project is pushing the ecosystem forward in the right direction. Like Kubernetes, Diego is a scheduler, and it operates mainly with cloud foundry. We also see Windows. NET integrating Diego from Uhuru, Docker has become a reality for windows, and I believe we will see these technologies converged.

Flynn

Flynn is an open source PAAs platform written in the Go language, Flynn uses modular design, and any module can be modified, upgraded, and replaced independently. The goal of Flynn is to simplify deployment and maintenance of applications in a distributed environment, and by using the git push command, Flynn can deploy applications to Docker, eliminating complex configurations and operations. Flynn is still in development, has not yet released a stable version, but has been funded by many companies, it is known as the next generation of open source PAAs platform.

Fig

Fig is a Docker based tool used to quickly build a development environment, and the FIG team has joined Docker company. Fig manages multiple Docker containers with a single configuration file, which is ideal for scenarios where multiple containers are used for development. Fig can work with Docker to build complex applications based on Docker, CoreOS is powerful but complex, and fig is relatively simple, but it is difficult to expand on multiple servers, and the importance of the project can be seen from the Docker official acquisition fig.