Rather than competing with cloud providers, enterprise IT departments should embrace the cloud business model.
The difference between enterprise IT-provided virtual machines and cloud-provided VMs comes down to complexity. Consumers have a much easier time monitoring, patching and updating cloud-provided VMs, thanks to consistency and standardization. Cloud customers buy fixed configurations from a provider's VM catalogue, a business model that many IT departments could and should adopt to improve efficiency.
What cloud providers provide
When you buy a
Additional reading on cloud business models
Choosing a less risky cloud business model
Viability of the SaaS business model
Why and when to adopt cloud infrastructure
That cloud-provided VM usually maintains the same configurations over its lifetime: An admin overprovisions it on day one, but never right-sizes the VM. In a cloud environment, there is often little to no redundancy in a single physical server or storage system, which makes those hardware resources less expensive on their own. The human expertise to optimize them, however, is much pricier. Cloud providers rely on virtualization platforms to optimize physical resource use and take advantage of low-cost physical resources, servers, storage and networking.
Many cloud providers also use free hypervisors to avoid incurring licensing costs as they add physical hosts. Service providers have always faced the risk of not earning money until adoption reaches a certain level. As a result, low acquisition and deployment costs often sit high on providers' lists of cloud architecture requirements.
Given the limited range of VM configurations, the cloud provider sees more workload consistency. Providers choose an overcommit level based on the service level agreement included in the product description. Cloud providers have no knowledge of or interest in the applications inside the VM; they simply deliver a service level as a degree of overcommitment for CPU, RAM, network and disk bandwidth.
What enterprise IT provides
In contrast, when an enterprise IT department provides a VM for a business unit, that VM is usually right-sized for the application and has adequate physical resources. As a highly engineered, highly redundant item, however, physical resource is costly.
Generally, enterprise IT uses a commercial hypervisor and applies redundancy technologies to anything that could fail -- from fans and power supplies to storage fabrics and storage area network controllers -- all of which drive up hardware costs. The effort and expertise required to tune and optimize VMs, however, ensures the best use of expensive physical resources. Enterprise organizations tend to accept operational expense if it helps avoid the capital expense of more hardware and licensing. When these organizations do purchase hardware, however, they will pay for the best.
VM catalogue shopping
Enterprise IT could very easily adopt cloud providers' practice of listing the available number of fixed VM configurations in online catalogues. Offering clear choices of VMs with defined standards will make both admins' and consumers' lives easier. Ideally, enterprise IT should also adopt the cloud provider's pay-per-use model with a chargeback system, or at least make business units aware of the computing resources they are consuming, a practice called showback.
Given the inevitable cost comparisons to cloud providers, such as Amazon, enterprise IT departments should show the value of everything they provide as part of the showback reports. Security and availability are expensive and often aren't included in a cloud provider's service. IT departments should also highlight redundancy and uptime, as well as place a value on backups and VM operating system patching and monitoring. Most cloud providers leave these responsibilities to the consumers.
Catalogues of easy-to-request services and VMs will ultimately help enterprise IT harness the efficiency associated with the cloud business model.
This was first published in May 2013