Manage Learn to apply best practices and optimize your operations.

Best practices for implementing a private cloud computing environment

Learn best practices for implementing private cloud computing environments and avoiding security problems in the public cloud.

When IT organizations realized that public clouds had some security issues, and they (IT) might lose control of their data center if the business app guys started using Amazon EC2, they quickly began to push for private cloud computing environments. Unfortunately, many IT managers equate private cloud with server virtualization. But they are not the same thing. Virtualizing servers is only part of implementing a private cloud. 

Some best practices for implementing private cloud environments:

  1. Do upfront planning --- Determine what a private cloud structure will do to your budget and financial flow. Chargeback should be an integral part of your private cloud computing environment. It makes users realize that they actually have to pay for resources used.  Determine the use cases for the cloud and create a strategy for how you will actually move apps to the private cloud. Which apps will run in the private cloud? Make decisions on apps running on Unix in proprietary clusters or on special hardware and mainframe legacy apps. Remember that the virtualization software runs on x86 servers.
  1. Select the implementers --- If you do your own implementation (not advised), you need good in-house expertise. You also have the option to contract professional services in private cloud vendors to do the implementation (very expensive), or hire a third party consultant to work with a private cloud vendor. Private cloud implementation is expensive and difficult.
  1. Know the implementation stages --- Start with server virtualization, follow that with storage virtualization, get virtualization-based management tools in place for monitoring, etc., implement a self service portal/service catalogue function, and add automation and orchestration tools. The order of the above actions can change, but you start with server virtualization.

    More resources on private cloud computing environments

    Demystifying the private cloud

    Adjusting IT attitudes toward private cloud

    Private cloud architecture: A blueprint

  1. Train employees/create new operational procedures --- The best way to train employees is to have them involved in implementing the private cloud. At least, have a select group of employees work closely with whoever is implementing the private cloud. You will have to create new operational processes for change management, provisioning, security, etc.
  1. Buy new software and hardware --- You will have to buy new software to monitor performance and traffic flow in the new virtualized environment. Traditional physical monitoring tools are not capable of monitoring traffic flow in the virtualized networks created by the hypervisor. You will almost certainly have to purchase some new hardware because some of the old hardware is not suitable for use with automation tools.
  1. Work with app business groups to satisfy their needs --- Private cloud computing environments involve two groups: IT operations staff and the business users who want to run applications. A properly implemented private cloud gives business app users the opportunity to quickly provision a server and run an application when they want to without human intervention. If this is not possible, then you need to rethink your private cloud implementation plans.
  1. Consider hybrid clouds for the future --- Plan for the possibility of integrating with a public cloud provider as you plan for and implement your private cloud environment. As you choose your private cloud vendor, pay attention to the virtual infrastructures implemented in public clouds to see how they might match up with the virtual infrastructure in your private cloud. This will help with cloud interoperability issues if you decide to move to a hybrid cloud environment in the future.

Dig Deeper on Public cloud and other cloud deployment models