Gregory Kloster

Exploiting the Economics of Pay-for-Use Cloud Computing

“In line with planned future investments in cloud computing technologies, a critical success factor will be establishing the unit cloud computing cost and unit value basis for any cloud service.”

by Gregory Kloster, July 2011

If today’s popular cloud-computing value proposition is a pay-for-use or utility-based economic model, why does there seem to be little investment by enterprise adopters or service providers in laying the foundations, tools, and methods for measuring and managing the economic efficiency of its cloud investments?

The hypothesis of this point of view is that cloud computing is a technology wave which triggers vendor investments in IT infrastructure and software. Factory enterprises lag in investing in technology and IT solutions for managing factory efficiency, capacity, and profitability.

Cloud computing as a business model represents an opportunity to embrace factory productivity economics. This is essentially the same method that factory owners use to calculate the unit cost of goods produced and the economic efficiency of their factories in response to scaling demand. These calculations typically use methods such as service-based costing analysis or cost volume profitability analysis. They are based on the foundation of unit cost management which enables metrics like unit cost and net margin contribution. The economic theory of the factory is based on assumptions such as the learning curve (the next unit of product costs less than the previous unit) and price elasticity (competitive changes in service or product price) to measure the economic efficiency of the factory. As a virtual factory, measuring cloud economic efficiency requires similar unit cost or service-based costing foundations as employed in measuring factory productivity and efficiency.

To profitably deliver a pay-for-use value proposition to customers, it should be straightforward for the enterprise or service provider to embrace “economy of scale” unit cost and net margin contribution benefits and the underlying foundations for business execution. But it isn’t.

Cloud computing and its economics represent an opportunity to deploy a pillar-like process for calculating its productivity, efficiency and profitability at scale. Achieving an extremely low-cost basis, exploiting the learning curve, and offering competitive pay-per-use constitutes business critical (and economic) measures of cloud performance. The economics of the “cloud factory model” should be a key feature by design in any investment in cloud computing. This hypothesis should be intrinsic to any implementation of public or private cloud infrastructure, platform, or application as a service.

In terms of strategy, the by-design principles which contribute to achieving cloud pay per use economic benefits include:

  • Service-based accounting and service catalogue that supports multiple cloud computing processes including unit cost Management, service management, operations, and IT asset life cycle management.

  • A methodology for scaling computing, storage, switching, bandwidth, and network capacity in terms of virtual machines, memory, ports, equipment upgrades, bandwidth, and cabling without service interruption.

  • At the infrastructure level, applying principles of modularity is key to scaling units of physical capacity. The industry standard for data centers stipulates a modularity standard that also serves to provide a methodology for scaling the physical cloud infrastructure.

  • Multi-layer instrumentation (physical through logical layers) to measure service performance and to calculate service unit cost and service unit cost value. Best practice is to build in instrumentation and measurement capabilities linked to unit cost management and capacity management tools.

  • Embracing dynamic elastic cloud computing management models and tools to effectively exploit spare capacity to achieve higher utilization factors and therefore lower service unit cost factors on a per transaction, per session, per subscriber, and per product basis.

  • A life cycle managed process that ensures 100 percent reconciled IT-asset physical and logical inventory and associated capital and operating expenditure cost allocation structure. This needs to be linked to any cloud computing resources and service catalogue.

  • Constructing a common information model, decision support, and information hub data warehouse capability for associating and correlating product service event classes (provisioning events, streams, sessions, transactions, utilization, resource consumption) to business rules for calculating and/or deriving service unit cost and net margin contribution. This should be integrated into an instrumentation and measurement methodology and unit cost management framework.

In line with planned future investments in cloud computing technologies, a critical success factor will be establishing the unit cloud computing cost and unit value basis for any cloud service. For example, just as factory enterprises measure profitability at the production and sales unit level, cloud providers that deliver pay-per-use pricing or chargeback models can reap similar benefits. These principles are aimed at marrying economic notions of price elasticity, productivity and economy of scale to enable a new paradigm for delivering IT value.

Gregory Kloster is director of Industry Insight and Strategy at Oracle.