As organizations increase their offerings of digital services, the demand for computing and storage capability increases. Also the need for faster, more complex data processing is becoming widespread. Additional demand for computing capability can only be met by increasing the processing capacity of servers within data centers. Along with greater computing capability, businesses have increased demand for storing digital data, both in terms of amount and duration due to new and existing applications and to regulations.
Just on time supply of these resources (processing power and storage capacity) can be an endeavor. Cloud computing can be an attractive solution to this. Cloud computing, using the definition of the National Institute of Standards and Technology (NIST) is a model for enabling convenient, on-demand network access to a shared pool of configurable computing resources (e.g., networks, servers, storage, applications, and services) that can be rapidly provisioned and released with minimal management effort or service provider interaction.
In economic terms cloud computing is offering:
- Elasticity; By dynamic (“on-demand”) provisioning of IT resources to customers, without customers having to worry for peak loads.
- Economy of scale; with multi-tenancy sharing of resources and costs across a large pool of customers by means of centralization of infrastructure and improve utilization and efficiency
- Shorter time to market/ time to create value; reduction in the average time to create and deploy a new solution and reduction in the average engineering effort to deploy.
- Different cost structure; shifting ‘from CAPEX to OPEX’ and pricing based on real consumption (utility computing).
And by this offering, cloud computing is improving business agility because of the ability to rapidly and inexpensively re-provision technological infrastructure resources.
With these compelling arguments what should stop you, in starting to use these cloud computing services?
In the paper “Do Clouds Compute? A Framework for Estimating the Value of Cloud Computing” by M. Klems et al. The authors are presenting a framework for comparing the two cases: Purchase a service from the cloud or buy hardware and set up your own IT infrastructure. The framework is based on a certain business scenario specified by the following parameters:
- business domain
- key business objectives
- demand behavior
- technical requirements derived from business objectives and technical behavior
But there is more at stake. As said by Steve Diamond, chair of the IEEE Cloud Computing Initiative, “Cloud computing today is very much akin to the nascent Internet – a disruptive technology and business model that is primed for explosive growth and rapid transformation.“ However, he warns that “without a flexible, common framework for interoperability, innovation could become stifled, leaving us with a siloed ecosystem.” (Therefore IEEE started two new standards development projects: IEEE P2301, Draft Guide for Cloud Portability and Interoperability Profiles; and IEEE P2302, Draft Standard for Intercloud Interoperability and Federation.)
And also from an economical perspective there are some items you should be aware of.
Who pays the bill of the IT infrastructure? Many stakeholders do not know much of the costs of IT and thus do not act on the IT infrastructure usage. This brings us to the concept of split incentives. Split incentives occur when the party that is responsible for paying the bills, is different from the party that is responsible for capital investment decisions. For example. A very obvious form of a data center split incentives case is that many data centers are housed in buildings that are not owned by the IT infrastructure user. If the user is not the sole tenant of a data center it is a shared service. The energy use per square meter of data center is allocated among all tenants rather than by actual energy consumption of each tenant. Tenants of leased data centers are generally unaware of or indifferent to the energy use of the IT infrastructure and have little incentive to make long-term investments in site infrastructure. Under this arrangement, most tenants never see the energy bill for their IT infrastructure. And then there is little to no incentive for the tenant to make a capital energy efficiency investment with a usual payback time of several years, and which in the end will revert to the site owner as property.
The issue of the split incentives problem can be seen as what is known as a principal-agent problem.
As Gravelle and Rees (Microeconomics, 2004) put it “a market is an institution in which individuals or firms exchange not just commodities, but the rights to use them in particular ways for particular amounts of time. [...] Markets are institutions which organise the exchange of control of commodities, where the nature of the control is defined by the property rights attached to the commodities”. If a party doesn’t have the right controls over the use of their commodities, exclusive use and/or the delegation of use and the related costs of doing so, this will result in inefficiency. An effect that is often visible in all kinds of sourcing arrangements of IT services.
Cloud computing efficiency improvements are burdened with discussions about costs. In particular, it is a discussion whether all the costs are taken into account for a proper costs evaluation. There is also a discussion about where to draw the line between transaction costs, switching costs, hidden costs or even production costs. This forms one of the barriers in the diffusion of efficient technologies. A proper overview of all the costs that are involved in diffusion of technologies could help to explain this barrier.
Production costs and upfront cost to obtain the efficient technology are not the only costs to take into account. Also, there is the barrier that the high upfront cost for efficient technology is more tangible than the money not spent on uncertain costs in the future. There are also transaction costs. According to the economist Ronald Coase (The firm, the market, and the law, 1992), transaction costs are resources that have to be used to carry out a market transaction, search, negotiation, verification, etc. Thus, this is the costs of arranging a contract ex ante where we have also the costs, ex post, to enforce the contract, as opposed to production costs. The transaction costs depend on the organizational set-up and the routines for making and implementing decisions. Traditionally there is the assumption that transaction costs to acquire and introduce new technology, are free. But in the different phases of a transaction, several sources of costs can be spotted:
- Planning; search for information, assessment of information, development of proposal, project identification and evaluation.
- Implementation; negotiation of contracts, procurement, project validation.
- Monitoring and verification; mechanisms to monitor, quantify and verify savings and related reductions.
There are also the switching costs. Switching costs are made when a customer makes a change of services and/or products. Types of switching costs include: exit fees, search costs, learning costs, cognitive effort, software and hardware costs, installation and start-up costs, costs for process and organizational changes, and financial risk. These costs are very much depending on the flexibility of the IT service stack: how the stack is technically build and the set-up of procedural routines and the organization of the stack. Hard interdependencies between the stack elements can put a heavy burden on substitution or replacement of service stack elements with better alternatives, because of high switching costs
in terms of money and effort. Factors that may form a barrier to switch include:
- High investment in non transferable infrastructure and/or software.
- Costs for changing proprietary interfaces.
- Costs for redefining configurations.
- Time to change, project duration.
These factors can even prevent or prohibit the supplier itself to change. System integrators or internal IT departments with large investments in data centers with accompanying complementary business models with ROI and NPV assumptions, can be very reluctant to make large changes to their infrastructure. The switching costs can even lead to a barrier in the form of a vendor lock-in, where the customer dependent on a supplier for services and products, is unable to use another supplier without high switching costs. A consequence of the cost discussions is that cost comparisons between several efficient solutions can be messy due to unequal comparisons.
Cloud computing isn’t just a discussion about technology, capex and opex it is about economics. Cloud computing is a disruptive new
business model with large consequences for the demand AND the supply side, with interesting micro-economical aspects (principal-agent problem).
For the customer as well for the supplier (external or internal IT organization) it all starts with a rethinking and proper comprehension of the business model that is being used. A method to describe a business model that has become extremely popular this last year is the business model canvas of Osterwalder (Business Model Generation, 2010). If you are unfamiliar with this concept have a look at this site or this slideshow Basically the concept introducing a standard language and format for talking about business models. Nine key items serve as the building blocks for all business models:
- Customer segments: Who will use the product?
- Value proposition: Why will they use the product?
- Channels: How will the product be delivered to the customers?
- Customer relationships: how will you develop and maintain contact with your customers in each segment?
- Revenue streams: How is revenue generated from which customer segments?
- Key activities: What are the key things that you need to do to create and deliver the product?
- Key resources: What assets are required to create and deliver the product?
- Key partners: Who will you want to partner with (e.g suppliers, outsourcing)
- Cost structure: What are the main sources of cost required to create and deliver the product?
These building blocks are laid out on a page (canvas) in a very specific way, referred to as a ‘business model canvas’. The business model canvas can be used to describe any of a wide variety of business models. With a proper understanding of the business model you want to achieve, you can rethink IT infrastructure ownership, rethink what are the controls you must have, and the change of command of IT infrastructure. This forms the starting point to define your cloud computing roadmap.
Call for writers
Cloud Computing technology has created new business opportunities but also created new questions. Question about ownership, make or buy, about sourcing. We started a new book project: ”The Sourcing Initiative: Enabling Collaboration“. The idea is that there are a lot of issues at stake in modern enterprises. Our proposition is that the modern enterprise must fundamentally rethink its sourcing equation (People, Organization, Economics and Technology) to become or remain viable. We have dubbed this Right Sourcing. It is about getting a good understanding of this equation and about finding a coherent set of ‘design‘
principles or a coherent way of working to make underpinned sourcing decisions and prevent wrong decisions. The book should be rich and diverse in content by taking different angles/perspectives by different writers. This is a non-profit project, and the book will be published under a Creative Commons licence and be freely available online. We will use an agile approach and release iterations as we go along.
We seek contributors who have something wise/clever/interesting to say about the themes and particular about cloud computing economics.
If you think you have and want to join us, please let us know. You can have a look at the website www.sourcing-it.org.