By Michael Nielsen, Director, HP Networking Solution Marketing
Five weeks. I am still reeling from hearing that statistic from one of our product advisors. The time to deploy applications is increasing—and in this increasingly complex environment, it takes up to five weeks to deploy a new service. Starting with the application, working through the VM profile, bare-metal availability, storage requirements, identifying the network characteristics and traffic profiles, staging, and cutover: it adds up to five weeks.
Why? Well, legacy networks were designed for connectivity and are indifferent to the applications that rely on it. They were, by design, rigid and inflexible. Many of them are still managed via CLI, or best case, a series of disparate management applications.
Recent advances in the industry have focused on solving discrete portions of the problem. Some vendors chased the data center to help speed server-to-server traffic and flatten the network so did we. Some vendors introduced separate, discrete architectures that were disjointed and failed to look at the problem across the entire infrastructure we didn’t.
The fundamental problem was a lack of unification of the solutions. Vendors weren’t looking at solving the deployment of the application, and the provisioning of the service from the data center out to the campus, branch, and mobile user.
Speed application deployment with Virtual Application Networks
This week, HP Networking announced our Virtual Application Networks strategy and the first phase of deployment. Our mission was simple: reduce the time it takes to deploy a new application from months to minutes, all the way from the data center to the end user, regardless of the application location.
Virtual Application Networks: a new way of thinking about IT service delivery
Virtual Application Networks are based on three guiding principles:
- Characterize the application to create consistency, reliability & repeatability across the entire network infrastructure
- Virtualize & program the infrastructure to create multitenant, on-demand, topology & device-independent provisioning
- Orchestrate based on templates, including user SLA and policy, to enable dynamic application delivery
Virtual Application Networks are logical networks, analogous to a virtual machine. They employ a control plane, which is analogous to a hypervisor. And they rely on a physical infrastructure, which is just like a bare-metal server. What I’m saying is this: the concept of virtualizing the infrastructure isn’t new, but it just hasn’t been applied to networking – until now.
Here’s how Virtual Application Networks works
With Virtual Application Networks, the server and network engineers decide on the network characteristics that are required by a virtual machine. Once that template has been created, it’s then applied into our Intelligent Management Center (IMC) as a Virtual Application Network template. Because we’re integrated with vSphere, we can see new VMs come online. As soon as they do, IMC immediately associates the virtual machine to the Virtual Application Network and deploys all necessary network configurations, instantly provisioning the entire network for that application.This can be repeated for all sorts of applications, optimizing a single network for potentially thousands of applications simultaneously.
We’re just getting started.
Virtual Application Networks is a journey, and our long-term plan will see us extend this functionality from the data center, out to the branch and campus, and to the remote and mobile user. How does that sound for your business? Let us know what you think and leave a comment below.
>> Explore related content
>> Learn more about FlexNetwork Architecture
>> Learn more about Virtual Application Networks
>> Learn more about OpenFlow