Container deployment
Container deployment of applications with flexible and efficient use of resources, the container can contain all the documents they need, like deploying applications on the same virtual machine can have its own configuration files and dependencies, but also has its own network interface. Therefore, running on a virtual machine applications like container application is easier than migrating applications installed directly, but also because each container applications are run independent of network interfaces, so it will not compete for the same the port in question.
Container at boot time, disk space usage, and CPU processing power has more advantages, because it is not running a separate operating system, it does not contain a large amount of software needed to run the entire operating system. It contains only application needed to run the software, and other tools want to run together with the container and a small amount of metadata describing the container. Container management tool is also more perfect, more mainstream management tools: Swarm, Kubernetes and Apache Mesos.
(1) Swarm is Docker native cluster tool, it uses standard Docker API, which means that the container can be used docker run command to start, Swarm will select the appropriate host computer running the container, this also means that other tools such as the use of Docker API Compose can also be used in the Swarm, which was to take advantage of the cluster instead of running on a single host.
(2) Kubemetes (often abbreviated to K8s) is Google's open source set of automated container management platform, formerly known as Borg, deployed for containers, automated scheduling and cluster management. Kubemetes currently has the following features: automated deployment vessel, automated capacity expansion or contraction, automation applications and service upgrades, container groups, provide services, support load balancing, health checking service, automatic restart.
(3) Apache Mesos island by A University of California, Berkeley. Lab first developed an open source cluster management software, support for Hadoop, Elasticsearch, Spark, Storm and Kafka and other application architectures.