Driven by fundamental changes in the economics, scalability, flexibility, and speed of implementation for computing, the cloud has started a transformation that will take place over a decade or more, not months or years. But thinking in the long term is challenging for CIOs and CTOs (CITOs) when most business are buffeted by rapidly changing market conditions.
In our view, planning for the cloud is not a matter of one upgrade or the implementation of a new set of tools here or there. Planning for the cloud will entail improving our understanding and management of IT.
Without a better understanding of what we already have, how it helps us, and what it costs, we cannot rationally decide what should stay on-premise, what should be in a private cloud, and what should be in a public cloud. Without admitting how little we know and planning experiments to know more, we will end up making huge mistakes.
This problem statement is focused on how CITOs can start to better understand both the cloud and their own computing landscape in order to build the foundation for a better business. We call the comprehensive vision for an ideal end state: The Open Cloud.
This problem statement is an initial attempt to understand the context facing IT professionals who are considering building a cloud and identify a roadmap in the short, medium, and long term toward a state in which the computing infrastructure is more valuable, more flexible, and more economical than it is today. The Open Cloud Visionaries series on Forbes.com is a set of interviews based on the ideas in this problem statement.
Here are the articles produced so far:
Know Thyself: The Fundamental Challenge of The Cloud The fundamental characteristics of the cloud are set forth in the popular definition of the cloud published by the National Institute of Standards and Technology, which defines the cloud as computing infrastructure as follows:
- On-demand self-service
- Broad network access
- Resource pooling
- Rapid elasticity
- Measured service
The definition then describes service models such as Software as a Service, Platform as a Service, and Infrastructure as a Service and deployment models such as private, community, public, and hybrid clouds.
In our view, too much attention has been spent on the mechanisms of the cloud not nearly enough has been spent on trying to determine what information is needed to make good decisions related to adopting the cloud. For example, let’s say you understand everything you can about cloud infrastructure and have a precise idea for how much it will cost to port an application to a public cloud.
Then a host of questions must be answered to make a decision to migrate an application:
- How much does it cost to run the application on-premise?
- How much will it cost to refactor the application to run in a public cloud?
- What value will come from the cloud mechanisms applied to the application once it is in the cloud?
- What are the implications of the choice of running the application in the cloud with respect to IT lifecycle issues such as security, disaster recover, service management, integration, and backup?
- What performance challenges will the application face based on the nature of cloud infrastructure?
- Will volume of data transfer become an issue?
This list is far from exhaustive. Some of these questions highlight weaknesses in current IT management practices. Others are focused on issues that will only be answered through experience. The point is that the cloud requires us to understand our own computing landscape better than we do now.
Another way to think of the transition to the cloud is to consider your current applications and then ask the following questions:
- What is the percentage of applications that are hosted on premise, in a private cloud, or in a public cloud?
- How will this percentage change over time?
- What applications will go where and why?
- What don’t you know about your applications that would help you determine the right home?
- What business benefits will accrue as part of the transition?
Again, the list is not exhaustive, and the questions are hard to answer. The answers will all come from better instrumentation and analysis of current applications and experimentation, not only from study of how cloud mechanisms work.
The ERP Analogy
Many ERP implementations were driven by the Y2K issue. If most companies considered the long term impact of their investment in ERP outside of Y2K, they would have thought more carefully about how to use ERP. If they were honest, they would have asked themselves: Do we really understand how ERP is going to change our businesses?
The answer then for ERP and now for the cloud is No. Of course, it would have been impossible to anticipate every way ERP would affect our businesses. But it would have been possible to recognize that ERP was going to have a long term impact that was difficult to understand.
Instead of rushing to implement it and then not really planning to keep learning about what works and what doesn't, many companies were exhausted after implementing ERP. It took years for the full impact and power of ERP to be understood.
There are now broadly two camps of companies who have implemented ERP. Those who find that its benefits are tempered by high costs and inflexibility. These companies don't leverage ERP to the fullest extent, but often feel bound by it.
The second camp is smaller. It is composed of the companies who understand their businesses in a detailed way and use ERP to run parts where it makes sense and other means where ERP is not a good fit. This camp often uses ERP as a way to standardize non-core activity and as a data collection mechanism to support custom applications and advanced analysis.
The cloud is now at a similar inflection point, but we should all be smarter now. We should recognize that we don't know the full impact of the cloud.
The first ways we use the cloud will only be the initial steps on a longer journey. The goal should be to rapidly experiment with the cloud to find the business impact. The result of this journey will be a new kind of computing, one that will deliver the long-standing promises of IT in a superior way for a lower cost in way that has a much wider impact.
There are a number of ways to organize thinking and activity related to cloud adoption given the context just explained. This problem statement takes the position that IT professionals should begin with the end in mind.
They should ask themselves:
- What are the long term implications of the decisions taken today?
- What is the process that we can use to determine what the cloud means to our business?
- How can we perform experiments to better understand our requirements?
- How can we lower the cost and risk of our experiments?
- How can we create and end result that provides our organization choice and power in the future?
Waiting is not an Option: What will Drive the Transformation to the Cloud?
The implication of the argument so far is that CITOs should pay at least as much attention to increasing their understanding of their existing application portfolio and infrastructure as to learning about the mechanisms of cloud technology.
In other words, in order to really make the most of the cloud, disciplines such as business process management, enterprise architecture, and IT Service Management all should be applied and improved. Newer methods for improving the understanding of the value of IT and the true costs such as Technology Business Management and IT Financial Management are another way to get the answers to the questions listed above.
But none of this will happen overnight, and the benefits of the cloud don’t have to wait until everything is completely figured out. That plan will fail.
What is needed is a set of experiments to better understand both the nature of the applications and the way that the cloud mechanisms provide opportunity.
The initial experiments in cloud adoption will likely be motivated by the following drivers:
- Savings achieved on various types of Infrastructure as a Service. This is essentially an extension of the victories of virtualization to the cloud
- Attractive Software as a Service applications will be adopted
- Software development will take place on Platform as a Service offerings for applications that can be hosted in the cloud
- For high value processes, the cloud may allow process redesigns that provide competitive advantage
- The cloud will be used to speed innovation and reduce time to market for key projects
- The cloud will be used to extend empowerment to end-users and developers but also the help manage the risks of that empowerment
Of these drivers, the first three are probably best understood, but are still in early stages of maturity for the most part. The last three are where the biggest value will come from, and the learning is just beginning.
The potential of the cloud is breathtaking to anyone with the slightest amount of imagination and ambition for technology to better serve a business. But the potential is not the same as a new working infrastructure based on the cloud. It is important to realized we are just beginning to learn about how to harness the power of this potential.
What Strategies Make Sense for Building an Open Cloud
So, to make good decisions about the cloud, know thyself. A full package of information needed to chart a path to an open cloud includes at least as much information about your existing infrastructure and its relevance to the business as it does about the cloud. That’s the first point.
The second point that should come shining through is that we are in the early stages of a long process. So much has changed in the past five years with respect to the cloud. So much more will change in the next five. So, one should be careful when choosing cloud infrastructure.
Here are the questions that should be considered when comparing cloud alternatives:
- How well suited is the cloud technology to your current needs?
- How mature is the technology?
- How much does it cost compared to alternatives?
- How mature is the ecosystem surrounding the technology?
- How well is the technology integrated into the current enterprise computing environment?
- How will the technology evolve to keep up with new developments?
- How will innovation be funded?
- What is the growth trajectory of the technology and its ecosystem?
- Who is in control of the technology? What are their motivations?
- What economic forces will determine the future price of the technology?
- What is the switching costs from the technology to alternatives?
To build an open cloud, you must not only look at the technology in its current form, but also as an evolving organism. The first few questions are about the current form of the technology and the last few address the way that the technology will likely evolve.
So, now the problem of the open cloud has been defined properly. An effort to build an open cloud will take place through the following stages:
- Through improved management of IT, a more comprehensive understanding of your computing infrastructure and applications will emerge. You will be able to answer questions about the cost and value provided that cannot be answered now.
- Through experimentation in using the cloud for infrastructure, as a development platform, and for Software as a Service applications, expected benefits will be confirmed and unexpected benefits will emerge.
- IT Management systems will be evolved so that on-premise, private, and public clouds can be managed in an integrated fashion.
- The analysis of cloud infrastructure will change from a focus on the current state of the technology to the analysis of the trajectory that the technology is likely to take.
- The focus of value created from the cloud will shift gradually from cost savings to business benefits derived from new kinds of processes and collaboration made possible by the cloud.
- The cloud will become the vehicle for empowerment of business units and individuals to create and manage their own IT solutions, and will also become the way that that empowerment is managed and made safe.
- Ultimately, the differentiation created by the cloud will be the speed of learning and innovation it makes possible.
In other words, building an open cloud will be the way that IT delivers more completely on the promise of IT to help a business thrive.