Mastering the Future of Cloud Computing
A Technology Leader
Before becoming chairman and CEO of VCE, Michael Capellas was chairman and CEO of First Data Corp., a global leader in electronic commerce and payment processing. Prior to that, he served as MCI’s president and CEO, where he oversaw its successful rebuilding and transformation from a long-distance carrier to a leading IP networks company.
Earlier, he was a key player in bringing about the merger of Compaq and HP. At Compaq, Capellas was successively CIO, COO, and then CEO, ending up as president of the merged HP. Prior to Compaq, he spent 20 years in technology positions for Oracle, SAP, and Schlumberger. Today, he also serves as a senior advisor to Kohlberg Kravis Roberts and First Data, and is on Cisco Systems’ board of directors.
Last autumn, CSC brought together business leaders to discuss the latest technology and security concerns. Michael Capellas, whose 30-year history of technology leadership has given him unique insights, led the conversation, discussing how the world has come to the brink of huge-scale technology-driven business change.
Capellas is chairman and CEO of VCE, The Virtual Computing Environment Company, a Cisco-EMC joint venture with investment from Intel and VMware, started in 2009 to bring together industry-leading IT infrastructure. In August 2010, CSC announced that VCE architecture will be our default platform for cloud-based deployments.
The following is an overview of Capellas’ views on what events and disruptions have led enterprises to the cloud and today’s IT landscape. Disruptive technologies, in particular the cloud, are likely to be adopted almost universally in the next few years: so how did we get to this point so quickly?
From leased lines to IP
Cloud’s story began in 2003, when most large organizations began to use Internet Protocol (IP) for their network management. Until then, most businesses used dedicated leased lines for data networking. The move to IP took place across the world in around 24 months; this fast switchover was all about business realities.
With an IP layer in place, network capacity can be dynamically assigned as needed, and users are only charged for what they use. They may no longer have to concern themselves with making predictions about usage and then committing themselves to renting a fixed amount of space: users can now assume that network capacity will be there when needed.
This marked a conceptual change for most large organizations, as they came to understand the commercial advantages of paying only for usage. Not only did costs generally come down, but it led to greater operational flexibility, with businesses no longer being constrained by availability, and instead able to pursue opportunity wherever it appeared. This provided confidence that their infrastructure would enable growth in any direction.
The processor revolution
The data communications revolution was followed in 2006 and 2007 with a step-change in processor power, with the catalyst being the arrival of Intel’s x86 processing architecture. Providing a combination of massive power in a very small space and greatly reduced power consumption, it lead to lower processing costs, and fast and fluid scaling.
This has begun to make the traditional microprocessordriven PC and server model obsolete. Computing needs can be managed through flexible groups of blade servers, which can be slotted into place as needed, leading to a new investment model and greater flexibility.
Expanding the edge
Finally, we are still seeing an explosion in wireless access. This has expanded the edge of both public and corporate networks, while also driving fast consumerization of technology, evidenced by the rapid adoption of smart phones and tablets. As millions of mini-apps spread across the world, barriers between previously separate technologies are breaking down.
Today, content is doubling every six months and includes voice telephony, video, animation, and business data. Content is also digital and can be created, shared, recycled, and blended within a world of multimedia applications and activities. These are the steps that have brought us to where we are now.
Reaching an inflection point
The cloud is the natural outcome of a decade of fast, evolutionary change. The new architecture reflects businesses’ demands for great flexibility, together with the explosion in content. It also means that networks can no longer be managed in a traditional way. But it takes something more than technology potential to create a true inflection point, and that extra factor is provided by economics.
The financial meltdown and the global recession that followed was an exceptional event, which occurs once every 50 to 60 years. All organizations need to reduce costs and share risks in order to survive, and that has driven demand for entirely new commercial models.
Significant breakthroughs also tend to occur immediately after periods of economic difficulty. Businesses are likely to freeze investment during the downturn, and when conditions change for the better, they want to make up for lost time by seeking big operational improvements.
That’s when they seek out and invest in radical new ideas.
A new approach to technology costs
Most businesses need commonsense reasons for moving to cloud-based service models, and the most compelling of these is that traditional models no longer represent value for money because:
- Businesses buy more software licenses than they need. They pay for nonessential applications because their legacy technology is too costly to switch off; they also guess at the volume of users, so tend to buy more than they can use.
- Data centers usually don’t run at capacity. They are filled with servers that are dedicated to specific applications, together with additional capacity for “emergencies.” Even a well-run corporate center may only use 30 percent of capacity.
- Spend on new applications is too low. New developments to support vital business activities rarely go above 20 percent of total IT spend. That means 80 percent of spend is used for “keeping the lights on,” rather than creating value.
That explains why return on IT investment is very low, which leads to problems in plentiful times. In hard times, this is simply unacceptable. It’s also now unnecessary.
Convergence in action
The cloud is the commercialization of converged technology, enabling organizations to divest themselves of many fixed costs and focus on their core business. For organizations, buying IT as a set of services, on demand, and paying only for what they need and use enables a revolutionary change for the better in their cost base and agility.
Most end users no longer think much about networking capacity because it is virtualized, thanks to the use of IP. We assume capacity will be there when we need it, and that we will only pay for what we use. Today, we can apply the same approach to a total infrastructure.
In this converged environment, end users do not need to worry about if they have enough servers or the right number of licenses. Applications and data storage will dynamically take up the space they need. It is an automatic, organic process.
Efficiency and utilization
The virtualized approach means that that an infrastructure can reach unheard-of utilization rates: close to 100 percent. And when demand seems likely to exceed supply, you simply buy another block of converged, combined processing, storage, and networking capacity, just as today you might slot in another low-cost blade.
The key principles are:
- No fixed provisioning, because in a virtualized environment, applications and data will find the capacity they need.
- No time spent on configuration, which may typically account for 20-to-25 percent of all IT department time.
- No long-term investment. You buy what you need and then simply scale up or down, as usage patterns flex with business demand.
The cloud is a means of automated virtualization and can be internal or external. Internal means buying combined processing and storage blocks, and taking more as demand grows. External means buying the infrastructure you need as a service. In either case, the result is greater simplification, greatly reduced investment levels, and dramatically reduced costs.
Lessons from the phone network
Capellas points to the global telephone network as a way of understanding how the cloud could work for us today. The international telephone network has been the world’s largest machine for almost a century. We can call anywhere in the world without wondering if the systems we are using will be compatible. The dial tone shows us the system is ready for business.
Telecommunication is much more complex than it was in the past, yet even as we invest in smartphones, which are basically specialized devices that transmit data and exchange multimedia content, we still expect the network to carry our traffic.
Just as we don’t need to know the details of the telephone network, increasingly users will cease to ask about the processing and storage that supports their business applications. It matters less and less all the time. What counts is the availability and integrity of the service, which will increasingly be sourced from specialists in the cloud.
The cloud stack
So what is the anatomy of the cloud? It is built from several distinct layers:
- Infrastructure as a Service: where the processing and storage exists and is made available to users, as needed, and on a scalable basis.
- Platform as a Service: a shared development and support environment that speeds up and simplifies the creation and management of business applications.
- Software as a Service: already familiar to many users through a range of commonly understood applications, this approach to software usage is replacing the traditional licensing model.
- Shared or federated clouds: enables users to place nonsensitive content or functions on a low-cost public cloud while keeping sensitive data inside their own firewall. This cuts costs and gives organizations the best of both worlds.
- Business Process Outsourcing as a Service: this represents a way of outsourcing complex processes on a pay-per-use, non-capex basis.
This is the landscape for the new, virtualized world of converged technology, in which applications, storage, and other technology and process-related requirements are delivered on demand as services. A cloud can be internal, external, or a combination of the two. The more of it brought in from specialist providers, the lower the costs are likely to be, but the technology framework stays much the same.