As the hype surrounding cloud computing begins to level out, the focus for businesses is no longer just on the flexibility and scalability of a cloud-based service but how to measure the effectiveness of such solutions, including ROI and service improvement. The demand for SaaS (Software as a Service) is driving the growth of cloud deployments, with independent analyst house Gartner, predicting that global spend on SaaS will rise 17.9 percent this year to $14.5 billion. According to Gartner, the SaaS market growth will remain strong through 2015, when spending on the software is expected to hit $22.1 billion. Much of the end-user demand for SaaS has resulted in the need for IT to implement cloud services, or at a minimum services that have the same attributes as cloud services. Many end-users are bypassing IT to directly access public cloud services, causing security concerns as well as productivity concerns as some needed legacy applications are not integrated with cloud applications.
As SaaS continues to evolve, and public and private cloud services gain traction, the utility computing aspect of service provisioning will increase the importance for the requirement to have more effective cost accounting and charge-back allocations. Let’s look at the attributes of cloud services that drive this change:
1. A catalog of services that are in demand by end-users
2. The capability for users to self-select desired services
3. Auto-provisioning so users have near-immediate access to those services
4. Accurate charge-back mechanisms to bill users for services used, and only those services
Looking at the above attributes, while cloud services hold a lot of promise, including cost savings, these characteristics in some ways create a ‘perfect storm’ for costs to escalate out of control.
The cloud, whether we are referring to public cloud services or IT-delivered private cloud services, makes it very easy for users to buy what they need, or want.
However, with this new capability, it becomes extremely important that the cloud services are ‘enabled’ in a way that ensures the business objectives of the cloud are met. In many cloud deployments, deploying the cloud technologies is simple and can be done in days or weeks. And in most of these cases even with the new cloud technology costs increase, users still do not adopt the IT-cloud services, and productivity is further degraded.
An important element of enabling a cloud to be effective is to design policies and procedures to ensure business goals are met. One aspect of this is establishing cloud-oriented cost allocations. As a user now equipped with the capability to select the services I want, and to have them immediately provisioned for me, several new cost-control elements need to be in place:
1. The cost for every service I select must be competitive, otherwise I will continue to bypass IT and go to public cloud providers to buy the services I need. As part of establishing competitive prices, IT has to allocate a fair market price, and cannot use the historical approach of assigning arbitrary corporate allocations to service users.
2. There must be a mechanism for decommissioning services that are no longer needed. If as a user I can easily provision new services, and pay for that provisioning, once I am done with the service unless some proactive mechanism takes that service out of commission a company will end up with a dramatically increased amount of idle services. As an example, users that can spin up new servers, use them temporarily and then forget about them will create massive server sprawl.
3. This decommissioning can be accomplished in part by policies, or in part by technology that is time-sensitive. It can also be controlled through cost allocations that charge a user not only for the initial provisioning, but for ongoing use or ‘in-service’ time.
IT organisations need to consider several factors when determining a competitive price for services.
Whilst IT cannot overly burden services with allocations to recover the cost of IT staff or the entire underlying infrastructure, reasonable costs for what is required to deliver the service can be included. Other factors such as data security and data sovereignty should also be considered. When looking to quantify usage, data centre managers need to assess a number of key areas including what data do you need to report on and does the relevancy of this data change depending on the customer? Quantifying usage is a grey area for many organisations with many still using standardised cost model, which does not work with a pay-for-use model. Another area to consider is whether the service being provided is being delivered to multiple devices i.e. mobile phones, PCs and tablets. Such considerations need to be taken into account in order to determine how to effectively charge-back while also meeting the organisations’ objectives. This task becomes challenging with heavily regulated sectors where organisations need to adhere with compliance standards, which can impact reporting requirements and data requests.
Deploying cloud technology is relatively easy. Ensuring that a robust, tailored cloud design, with clear business objectives and redefined procedures, policies and methodologies is in place is one of the biggest challenges for CIOs. A vendor independent, objective data centre and managed services consultancy can provide this cloud enablement. The resulting success of a cloud deployment can have a positive impact on a product or service’s time to market, and the business will recognise immediate productivity gains. Most importantly, this will provide a tremendous step forward in positioning IT as a valued partner and service provider to end users.
Tags: Private Cloud, Public Cloud, Hybrid Cloud, Cloud Security, Service Providers, Software-as-a-Service