Preparing for Your Private Cloud, Part III: Decide How You Will Provision and Manage
Standard management practices shouldn’t go away when you adopt a private cloud. You should have a way to measure, monitor, and improve its effectiveness. Security policies must be maintained – in fact these policies may need to be strengthened to accommodate the pace at which cloud environments are adopted and provisioned. Else you may risk reputation, customer relationships and regulatory compliance. As evidenced by several analyst reports this year, without good virtualization management maturity your private cloud may end up costing more than traditional infrastructure.
This means you’ll need a combination of standard configurations, tested tools and effective policies to succeed with private cloud. In Infrastructure as a Service (IaaS) environments, for example, you should
- Establish multiple standard selections for users who will be requesting access to a new virtual machine (VM). This might be a selection of a small VM, a medium configuration, or a large configuration. These configurations should also allow a choice of operating system, disk sizing/scaling and network connectivity options.
- Decide what user roles are allowed to provision what size and type of environments. Administrative users might only be allowed a small configuration, knowledge workers a medium configuration, and programmers or testers a large configuration.
- Configure the software and networking in these environments to comply with security and organizational policies.
- Determine the lifecycle for VMs and the conditions and timing under which they will be shut down to recapture cloud resources.
- Make these configurations standard selections for self-provisioning through a self-service portal – this will let you focus systems administrators on other tasks and improve IT efficiency.
- Any requests outside of these standard configurations should be automatically routed using workflow tools to speed getting the request to the right authorizers quickly and automatically. This can improve worker satisfaction and increase business agility.
A private cloud will strain your service management processes – for example, the volume of change requests will increase enormously with self-provisioning of VMs, putting a strain on both your Change and Configuration Management processes. Many more incident calls may hit the Service Desk. So pay attention to core IT processes such as Incident, Problem, Change, Release and Configuration Management and assess the impacts your new cloud will have on these processes. Decide what monitoring, analysis and reporting tools you will use to identify incident and capacity trends, monitor SLA compliance, and detect security breaches.
Answer hard questions such as: How will the Service Desk restore normal service and how will Problem Management perform root-cause analysis if application instances move too quickly to be registered in a CMDB (Configuration Management Database)? What Change Management policies are in place to ensure the right pre-approvals are used to allow smooth and rapid provisioning, while preventing unwanted changes from creeping (or roaring) into the IT landscape?
How will you charge equitably for actual use of the shared infrastructure? Chargeback technology will be of little value if not coupled with a sound IT Financial Management process and policy. Getting everyone to agree on a shared financial model may be your biggest challenge.
As you can see, building a private cloud is about much more than just the technology! So before you turn the lights on your private cloud, make sure you can measure, monitor and secure to ensure the effectiveness of your shiny new cloud.
In my next blog post, I’ll talk about IT industrialization and creating what I like to call a cloud factory.
Tags- Cloud computing Infrastructure as a service Security Service management
" > 
" > 
" > 
" >