But realizing the benefits of fundamental data center changes requires a series of best-practices decisions along the way. When implementing a private cloud strategy, it is crucial that CIOs and IT managers consider the changes that will inevitably take place in the data center as the vision of their IT infrastructure takes shape.
The key is to start by identifying the business challenges to be met. Then examine the technology assets in place and the companys financial goals and objectives. There will be pros and cons to every decision, and the effects on the data center and its ability to deliver price-performance to the company must be taken into account at every turn.
Here's a review of the five major changes a private cloud will bring to your data center:
1. Lower cost - Through an effective virtualization strategy and a resultant reduction in servers, data centers can reduce the amount of cabling, floor space, and power they consume by record amounts.
On the flip side, virtualization is only the beginning of a journey into the cloud; systems also need to be able to dynamically provision IT resources as needed on-the-fly.
Less control - Instead of having to over-provision (and, therefore, spend) to meet changing demands, the addition of a self service portal above the orchestration level provides on-demand consumption. This enables the IT team and designated end users to spin up capacity as needed thereby increasing business agility. Caveat emptor: The idea of giving end users easy access to resource control leaves CIOs and other IT professionals nervously pacing the floor.
Increased need for automation - Imagine that a functional cloud environment looks something like an onion with traditional infrastructure at the center and a self service layer around the outside. Between the resources and the self service portal, there must be distinct layers for automation, monitoring and governance.
Without automation, any savings gained through virtualization and self provisioning is lost on the additional IT staff needed because there is no "staff-on-demand" option that will take effect just when you need it. Without monitoring and management tools in place, self service provisioning could seem like the Wild West all over again.
Power and cooling - While a private cloud is responding to changing business demands on the front end, the back end of the cloud needs to be able to respond to changing demands for power and cooling. Servers that become virtualized require specific power and cooling needs. For example, as private cloud features like self provisioning are introduced, hot aisles may get hotter requiring more cooling in that region.
Power consumption fluctuates when virtual machines are spun up and down and will require smart power distribution units (PDUs) that can monitor power and respond to changing demands. Data center managers should use computational fluid dynamics (CFD) modeling to identify hot and cold spots in their data centers. By monitoring CFD results, data center pros can see where airflow and cooling inefficiencies exist and model improvements.
New visibility requirements - With virtualization, self provisioning and other cloud enabled advances, it is now possible to develop applications that can "right-size" themselves dynamically according to pre-determined criteria.
Thats great for new applications, and great news for IT departments.
But there are still decisions that need to be made by, dare I say it, a live person. So monitoring the data center in a private cloud environment is key. Data center infrastructure management (DCIM) tools should be deployed so the operations teams can actually see what is happening in the data center with regards to these changing computing capacity, power and cooling dynamics.
Private clouds leverage the equipment and skills a company already has in its data center. This makes them extremely cost effective solutions for delivering IT as a service. But its also important to recognize that new tools may be needed to take full advantage of everything a private cloud has to offer.
Automation and orchestration tools, properly implemented, are the key to moving from a virtualized environment to a private cloud. And knowing just what needs to be added or changed in the data center can only come from a thorough assessment of both the companys business needs and ITs ability to meet those needs.
Kevin Gruneisen is vice president of data center solutions for Logicalis, an international provider of integrated information and communications technology (ICT) solutions and services.