Please enable JavaScript.

Coggle requires JavaScript to display documents.

Containers (Containers and Clusters for Edge Cloud Architectures – a…

- - - - The current third generation of PaaS includes platforms
        like Dawn, Deis, Flynn, Octohost and Tsuru,
        which are built on Docker from scratch and are
        deployable on own servers or on public IaaS clouds.
  - - - Though, a TOSCA blueprint can also be used
        for the basic container case to, for example, spawn a number
        of instances of a Docker image.
      - Currently, TOSCA deals with container technologies, such
        as Docker and Rocket, as well as container management technologies,
        such as Kubernetes and Mesos, in an agnostic way.
        
        Due to this abstraction and interoperability, TOSCA is a suitable
        platform for defining a standard container orchestration
        specification that is portable across various cloud environments
        and container providers
- - - - The process by which it gets in front of a customer, that workflow of dev test, staging, and deployment to production, is one of the most [tension-fraught] in IT. (p. 102)
- - - - Docker containers are created using a base image and running a Dockerfile on that image.
        
        Dockerfiles are scripts composed of various commands and arguments listed successively to automatically perform actions on a base image.
  - - - Adopted by Google Cloud Engine, Microsoft Azure, VMware and others.
      - (this article was published in 09/2014, before the release of Docker Swarm in 11/2015)
- - - - Virtual Machine Monitor (VMM)/Hypervisor
      - Virtual Machine (VM)
        
        A self-contained operating environment that behaves as if it is a separate computer.
  - - - It builds a cooperative group of systems that can provide redundancy of one or more nodes fail.
  - - - Dockerisation consumed fewer resources and operational overheads compared to virtualisation.
      - The dockerisation-based distributed system utilised less memory, storage, CPU usage, CPU time, boot time and energy within the same Linux-based environment.
      - Docker Swarm has inbuilt features of scheduling, load balancing, discovery, availability and networking that help it to act as a natural distributed system.
        
        VMs are completely independent from each other, so require extra/external provision to coordinate and act as a distributed system similar to Docker Swarm.
      - In the future, it may be interesting to simulate and evaluate dockerisation and virtualisation technologies in the cloud.
- - - - It has accepted significant
        numbers of pull requests from the likes of Red Hat, IBM, Microsoft, Google, Cisco
        and VMware. (p. XX)
- - - - Application services can be scaled across nodes by creating and managing several containers from the same image.
        
        Orchestration still faces challenges.
        
        (this article was published in 05/2015, before the release of Docker Swarm in 11/2015)
- - - - Our tests used three widely recognized Hadoop benchmarks: TestDFSIO, MRBench, and TeraSort. They were speciﬁcally chosen to compare the system performance of V-Hadoop against the two Hadoop modes with respect to different resources: TestDFSIO for disk intensive jobs, MRBench for CPU intensive jobs, and TeraSort for CPU, disk and network intensive jobs. (p. 239)
        
        The prototype version of our V-Hadoop framework can manage a V-Hadoop cluster across multiple physical machines, and can perform simple, resource-based scheduling of Linux containers across physical machines by adding or removing containers in the cluster dynamically based on resource usage and availability. (p. 241)
        
        Our experimental evaluation shows that VHadoop performs comparably to a fully-distributed Hadoop cluster. V-Hadoop allows for elastic clusters that can utilize the resources of the underlying physical infrastructure, providing signiﬁcant management and cost beneﬁts to the user. (p. 241)