by Paul Rudo on 26/05/12 at 11:44 am
OK. So you’ve decided to move your servers to this magical place called “the cloud”.
But wait! Have you thought about where your cloud will be hosted? Choosing the right cloud topology is one of the most important decisions you can make, because it will impact the costs, security, and control of your cloud server deployment.
That’s why I’ve decided to lay out a few of the most common options you may run into:
This is the most common and most cost-effective form of cloud server hosting.
In a public cloud, all of the hardware and infrastructure is owned and managed by a third-party. With this approach, you only pay for computing resources without worrying about any of the behind-the-scenes hardware.
Dedicated Private Cloud
With a dedicated private cloud, you host all of your servers on virtual machines that reside on hardware which is owned and controlled by you. And this hardware is stored inside of a datacenter which resides within your internal network.
This is the most expensive type of cloud-computing infrastructure. Because of the costs involved, it’s usually more suitable for large organizations, or companies with special security and privacy requirements. (Although the costs have been falling, and virtualization hardware is becoming much more accessible)
Managed Private Cloud
With a managed private cloud, you own and control all of the hardware, which is usually hosted at an external site where it is managed by a third party. This way, you get most of the security and control benefits of a dedicated private cloud, but without having to build a new datacenter or hire any new staff.
Hybrid clouds give you access to the cost-saving benefits of public cloud computing, while also offering the security and control of a private cloud.
How does it accomplish this?
Within every company, there are large amounts of information with varying amounts of sensitivity. So you can achieve cost-savings by only using the private cloud for highly confidential computing tasks, while sending non-sensitive or non-critical tasks out to the more cost-effective public cloud.
A good example would be a medical research company that removes personally identifiable information from their records before sending it off to a supercomputer for advanced analysis.
Of course, there are other variations on the cloud topology theme, but most of them can broadly fit into one or more of the topologies I’ve outlined above.
Has your organization implemented a public or hybrid cloud topology? Leave a comment below and share your experience.