How Private Cloud Differs from Traditional on-premise Computing ?

Private Cloud vs Traditional HardwareMany users describe private cloud as having an Amazon AWS-like environment within their own data center. This is where software developers/engineers can easily and quickly request and receive application infrastructure, just like they can on Amazon AWS. The difference is that the underlying servers, storage and network is wholly owned and dedicated to the organization instead of rented through AWS or other public cloud vendors. A private cloud can be within the organization’s own physical data center or owned, dedicated (not shared like public cloud) and managed by a third party.

Compared to a traditional computing infrastructure, private clouds require five main components: IaaS software (i.e., VMware vCloud, Microsoft System Center, OpenStack), PaaS software (i.e., Cloud Foundry, OpenShift), virtualization software (i.e., VMware vSphere, Microsoft HyperV, Xen, KVM), automation / configuration and orchestration software (i.e., Puppet, Chef, and Ansible which was acquired by Red Hat), and dashboard and management tools (i.e., VMware, IBM, Microsoft).

Private clouds have five main attributes: self-service, a dedicated environment, cost allocation, speed and uniformity, and scale. This is different from current traditional on-premise environments because software developers can now acquire the infrastructure they need in minutes or hours instead of days or weeks without involving IT. That infrastructure is uniform and can be easily managed and scaled. In addition, the customer has control over how they deploy their workloads and manage their cluster, and can make sure that their environments meet regulatory requirements.

Similar to public cloud, the most common use cases cited are new applications (web, mobile, IoT) and development and test workloads. Some applications were in production, and some migrated from traditional on premise to private cloud.
Typically organizations build a private cloud (versus public cloud) due to regulatory concerns or wanting more control over how their cloud runs. Examples of this include AT&T, PayPal, Walmart Labs, Intel, FICO, Volvo, etc

Overall, we expect that the overall spilt between platforms will be roughly 50% public cloud, 20% private cloud and 30% traditional over the next 5-10 years, similar to Cisco’s Global Cloud Index forecast. This is based on commentary from companies like General Electric, which plans “to move 60% global workload into AWS,” Coca-Cola,
which expects 50% will run in the public cloud, and Johnson & Johnson, which expects 85% of its workloads will be in a private cloud, and 15% will remain legacy.