Last week I met with a Service Provider to continue some discussions that we have been having about the adoption of VCE’s Vblock into the reference architecture for their managed Private Cloud service. I thought I’d write a quick post reflecting on the discussions as there were a couple of interesting points that were raised. No names I’m afraid as I need to respect a client confidentiality agreement.
For the skeptics, it’s worth stating that there are many service providers delivering services of this type today: Dedicated, (largely) virtualised infrastructures with elastic properties and self service management interfaces, delivered either on the customer’s premises or in a Service Provider data centre.
- A delivering a reduction in cycle time to provision new customer infrastructure, or to expand existing infrastructure
- Reducing the amount and variety of skilled resources required to provision infrastructure
Firstly, let’s look at the reduction in solution deployment cycle time. It might not be immediately obvious unless you’ve worked within a Service Provider, but a significant amount of effort is expended in managing the design, purchasing, logistics, assembly, integration and testing of even a moderately sized network, storage and computing infrastructure. Service Providers can offload these activities to VCE if they chose to standardise on the Vblock platform. This can free up resources to work on other value adding activities.
Provisioning of physical resources within a Vblock is automated using EMC’s Unified Infrastructure Manager (UIM). This tool enables compute, storage and network resources to be provisioned via a simple web based interface, without the operator needing to have detailed knowledge of any of the underlying element management or command line interfaces that would be required to carry out these activities manually. This means that generalist or even design resources can directly provision physical resources within the Vblock and present these directly to VMware for subsequent provisioning of virtual resources, or to an operating system deployment tool, should they be required for a bare-metal OS installation.
Of course, these benefits apply equally in an Enterprise IT environment, the only difference is that in a Service Provider, the impact of sub-optimial approaches to infrastructure deployment and operations is magnified due to the sheer scale of the environments that exist there!
Update (22/Dec/11): Since writing this, I have been doing some reading around the subject and I came across a new piece of research from IDC entitled ‘Converged Infrastructure Survey 2011: Out of the Clouds, Down to Earth’, which you can find here (requires IDC subscription). Reading this, it’s very interesting to note that the top three reasons that enterprises cite for making a decision to adopt converged infrastructure are:
- Improved availability
- Simplified management
- Lower total cost of ownership (TCO)