Quite a few people have been asking me to write this article for a while. vCloud Director has been around a while now, but it appears that it is only now gathering momentum and it is being looked at seriously. This multi-part article will cover all the basic concepts of vCloud Director and what the different terminology means.
First of all, lets talk about what VMware vCloud Directors actual purpose is:
A quote from the VMware vCloud Director page on VMware.com explains:
Manage resources more efficiently by logically pooling infrastructure capacity into policy-based virtual datacenters. VMware vCloud Director integrates with existing VMware vSphere deployments and extends capabilities like Distributed Resource Scheduler (DRS) and vNetwork Distributed Switch, to provide elastic compute, storage and networking interfaces across multiple clusters. Using virtual datacenters built on top of vSphere, VMware vCloud Director enable resources to be provisioned without the need for repeated configuration or significant maintenance.
The way I like to explain vCloud Director is this way:
vCloud Director is an abstraction layer, it allows anyone to come along and consume compute, memory, networking and storage resources without worrying about the underlying hardware or infrastructure. It enables that current buzz word Infrastructure as a Service (IaaS). Along with providing abstraction, it also allows consumers to have an on-demand self service portal allowing them to provision there workloads when and how they want them. Now people often comment: well that is just orchestration. This is not orchestration per say, but more policy based provisioning allowing the consumer to take more control and ownership of there workloads.
There are more articles out there that discuss what Cloud actually is, it is not my intent to do this, I am merely providing an overview of the product.
So we will start with what is a consumer? There are two terms that VMware use when talking about vCloud, that is Provider and Consumer. These are two very different concepts, but do have some cross over points.
The easiest way to look at this is a Provider is your Cloud Admin or Service Provider, they are the ones who care about the underlying hardware and infrastructure, they care about capacity planning and making sure the virtualisation layer is performing as expected. A consumer on the other hand, doesn’t care (and yes I know some people disagree and think they should care). They are paying for a service with an expected SLA, so they just want to run there workloads without thinking about what version of vSphere am I running, or what storage array should I deploy my VMs on. They just simply want to deploy workloads in the form of Virtual Machines and pay for the resources. This is the case both for Public Clouds and Private Clouds. Enterprise IT departments are no different to a Service Provider. They are simply offering a service to an internal departmental cost centre, rather than a public consumer.
Business Requirements for Cloud
Now I guess most people already understand key business requirements from a Service Provider perspective, it is to provide a service, whether that be Infrastructure as a Service, Platform as a Service, or even a Managed Service. However, why would an Enterprise or Medium sized business be looking at having an on-premise private cloud? The list below demonstrates some of the business requirements that I have spoke about previously at VMworld and some VMUGs
- Innovation and new product development
- Accelerating release cycles and speed to market
- Prolonging legacy applications
- Operational efficiency
- Reduced TCO
- Business Agility:
- Multi-tenancy support
- Self-Service capabilities
- Designed for Scalability and Elasticity
- Metering capabilities for cost reporting
- Leverage shared infrastructure and resource pooling
- Provide differentiated offerings based on cost
Some of the key takeaways from this list is Operational Efficiency and Reduced Total Cost of Ownership (TCO).You will be saving money on buying hardware and running costs (power/cooling/maintenance) and Operationally staff will be able to concentrate more on project based work rather than break/fix all the time. One other favourite of mine is the ability to prolong legacy applications, how many people in an Enterprise have development teams with servers under there desks who refuse to relinquish control of these? I once read an article that described these people as Server Huggers and it has always stuck with me. You now have the ability to virtualise these servers and give full control of the management to the individual themselves. This in itself reduces the worry and management complexity of trying to support these machines.
In Part two of this article I will be discussing the VMware vCloud Director concepts and start to look at the technical aspects of Virtual Data centers.
I’m just starting to look at vCD so this series is welcome. Looking forward to future parts.
really a clear and useful expression v vCloud Director
Very much liked the post.Waiting for next!!
Very concise thanks. Please consider including a link to later parts. For instance from part 1 to part 2.
Thanks for the tip, updated the post.
Great article, David. I do however have one helpful suggestion on improving your grammar and making your posts even more excellent – please take the time to identify the differences between ‘there’ and ‘their’. In your article, there are many instances where the latter should have been used instead of the former.
Specifically, ‘there workloads’ and ‘there desks’ are the instances where ‘their’ should be used.