By Adam Ronthal
The main motivations for moving data management to the cloud have always been cost savings and operational efficiencies. Data resides in, and is processed in, the unlimited cloud space, eliminating the need for expensive on-premises data centers with high maintenance fees and limited capacity.
While that’s been the common belief for a long time, it’s looking as if the reality may not coincide with this thinking. The majority of respondents to a 2016 Gartner survey said that using the cloud had actually increased their operational expenditure.
Data and analytics leaders who use cloud infrastructure for analytics, data lakes, and other data-intensive environments are often surprised at the size of their monthly cloud spending. The key reason for this is that on-premises data centers and the cloud work according to different economic principles. On-premises, you tend to maximize utilization, as capacity and cost are fixed. In the cloud, it’s opposite: Users will be charged for every single computing action, so they must aim to consume as few resources as possible.
This doesn’t mean you should spurn the cloud altogether — there are options for cost optimization in the cloud that don’t exist on-premises. As will be discussed during the upcoming Gartner IT Symposium/Xpo in Toronto, data and analytics leaders must apply financial governance controls to their cloud environments.
Most cloud service providers (CSPs) already provide options to apply budgets and quotas at both account and individual-project level. There are native budgeting and forecasting services with varying degrees of integration for existing data management solutions. Data and analytics leaders should educate themselves about the different options and set budgets for their various cloud activities.
Proactive budget controls are the first step toward managing cloud budgets and enforcing predictable spending. However, native controls are often very basic, as CSPs focus their attention on services that offer them more value.
After you set budgets, decide what to do when budgetary capacity is exceeded or unexpected spending occurs. Options range from sending a notification to the resource owner to more drastic steps like throttling back capacity until the next budget cycle or blocking access to the exceeded resource.
In addition to budgeting and forecasting capabilities, many CSPs provide features at the service offering level that enable finer control of resource allocation — and consequently of the cost of resources. An example is the separation of compute and storage resources; if users can independently scale these, there are many ways to realize cost savings.
Much of the integrated functionality for tracking and budgeting within CSP frameworks is still in its adolescence, and often requires significant customization. Independent software vendors (ISVs) have seized the opportunity that this relative immaturity presents by building additional capabilities for financial governance directly into their cloud-based offerings.
ISVs, in general, have a more immediate interest to build cost optimization and financial governance controls than CSPs do. This accounts for some of the more advanced tools in some ISVs’ data management offerings.
Adam Ronthal is a Research Director in Gartner’s IT Leaders (ITL) Data and Analytics group with a primary focus on database management systems, technologies and strategies. Mr. Ronthal’s areas of specialization include both operational- and data-warehousing-focused use cases, cloud, and traditional implementations.