Linux containers, as a lighter virtualization alternative to virtual machines, are gaining momentum. The High Performance Computing (HPC) community is eyeing Linux containers with interest, hoping that they can provide the isolation and configurability of Virtual Machines, but without the performance penalties.
In this article, I will show a simple example of libvirt-based container configuration in which I assign the container one of the ultra-low latency (usNIC) enabled Ethernet interfaces available in the host. This allows bare-metal performance of HPC applications, but within the confines of a Linux container.
Read More »
Tags: HPC, Linux, Linux Containers, USNIC
I find Linux containers among the most fascinating technology trends of recent past. Containers couple lightweight, high performance isolation and security with the ability to easily package services and deploy them in a flexible and scalable way. Many companies find these value-props compelling enough to build, manage and deploy enterprise applications. Adding further momentum to container adoption is Docker, a popular open source platform for addressing key requirements of Linux container deployment, performance and management. If you are into historical parallels, I can equate the Docker evolution and growth to the Java programing language which brought in its wake the promise of “write once run everywhere”. Docker containers bring the powerful capability of “build once and run everywhere”. It is therefore not surprising to see a vibrant eco-system being built up around Docker.
The purpose if this blog is to discuss the close alignment between Cisco ACI and containers. Much like containers, Cisco ACI provides accelerated application deployment with scale and security. In doing so, Cisco ACI seamlessly brings together applications across virtual machines (VM), bare-metal servers and containers.
Let us take a closer look at how Containers address issues associated with hypervisor based virtualization. Hypervisor based virtualization has been a dominant technology in past two decades, with compelling ROI via server consolidation. However, it is well known that hypervisors bring workload dependent overheads while replicating native hardware behaviors. Furthermore, one needs to consider application portability considerations when dealing with hypervisors.
Linux containers, on the other hand, provide self-contained execution environments and isolate applications using primitives such as namespaces and control groups (cgroups). These primitives provide the ability to run multiple environments on a Linux host with strong isolation between them, while bringing efficiency and flexibility. An architectural illustration of Hypervisor based and Container based virtualization is worth a quick glance. It is apparent from below, Docker based containers bring portability across hosts, versioning and reuse. No discussion on Docker containers is complete without mention of DevOps benefits. Docker framework – altogether with Vagrant, for instance – aligns tightly with DevOps practices. With Docker, developers can focus on their code without concerning about the side effects of running it in production. Operations teams can treat the entire container as a separate entity while managing deployments.
ACI and Containers
Cisco Application Centric Infrastructure (ACI) offers a common policy model for managing IT applications across the entire Data Center infrastructure. ACI is agnostic to the form-factors on which applications are deployed. ACI supports bare-metal servers, Virtual machines and containers, and its native portability makes it a natural fit with Containers. Besides, ACI’s unified policy language offers customers a consistent security model regardless of how the application is deployed. With ACI, workloads running in existing bare-metal and VM environments can seamlessly integrate and/or migrate to a Container environment.
The consistency of ACI’s policy model is striking. In ACI, policies are applied across End Point groups (EPG) which are abstractions of network end points. The end points can be bare-metal servers, VMs or Containers. As a result of this flexibility, ACI can apply policies across a diverse infrastructure that includes Linux Containers. I want to draw attention to the ACI flexible policy model applied to an application workload spanning bare-metal servers, VMs and Docker containers as illustrated below.
You may recall Cisco announced the broad endorsement for OpFlex protocol at Interop Vegas 2014. We are currently working on integrating OpFlex, Open vSwitch (OVS) with ACI to enforce policies across VMs and Containers in earlier part of next calendar year.
As Container adoption matures, managing large number of them at scale becomes critical. Many Open source initiatives are actively working on scalability, scheduling and resource management of containers. OpenStack, Mesos, Kubernetes are among the open source initiatives / communities Cisco is actively engaged in to advance ACI integration with open source tools and solutions.
With containers, we have seen only the tip of the iceberg. Docker containers are beginning to get traction in private clouds and traditional Data centers. Cisco ACI plays a pivotal role in integrating ACI unified policy model across a diverse infrastructure comprising bare-metal, VMs and Containers.
For more information refer:
Tags: ACI Policy Model, bare metal, Cisco ACI, Cisco APIC, docker, Linux Containers, opflex protocol, virtual machines
Reduction in the complexity of deploying and managing services, accelerating new service introduction, and reducing capital/operational expenditure overhead are key priorities for network operators today. These priorities are in part driven by the need to generate more revenue per user. But competitive pressures and increasing demand from consumers are also pushing them to experiment with new and innovative services. These services may require unique capabilities that are specific to a given network operator and in addition may require the ability to tailor service characteristics on a per-consumer basis. This evolved service delivery paradigm mandates that the network operator have the ability to integrate policy enforcement alongside the deployment of services, applications, and content, while maintaining optimal use of available network capacity and resources. Read More »
Tags: architect, capacity planning, Cisco, decoupling, delivery of services, deployment, engineers, extensibility, infrastructure, innovative services, Linux Containers, network topology, onePK, resource pooling, Servers, service appliances, service delivery, services, technology