Ever since cloud computing first appeared in the IT world, people have struggled to define what it is and why it’s so important. “There is a clear consensus that there is no real consensus on what cloud computing is,” said the organizer of a 2008 cloud conference in his closing remarks. Just about everyone agreed that something big and profound was going on, although we were not totally sure what it was yet.
Is cloud the 21st century version of time sharing, enabling users to get virtual access to different kinds of IT resources and software capabilities without having to own computers? Is it a return to centralized computing, driven by the rising complexities, management costs and energy inefficiencies of distributed systems? Is it the evolution to utility computing, where, as with electricity and water, you get your IT from large service providers and pay for it based on usage?
Or, as The Economist said in an excellent 2008 article, is cloud the next step in the brief history of computers, as the machines “are evaporating altogether and becoming accessible from anywhere”. . . as computing is becoming “more and more disembodied and will be consumed where and when it is needed.”
“The rise of the cloud is more than just another platform shift that gets geeks excited,” it presciently added. “It will undoubtedly transform the information technology (IT) industry, but it will also profoundly change the way people work and companies operate. It will allow digital technology to penetrate every nook and cranny of the economy and of society, creating some tricky political problems along the way.”
I like the way The Economist put it. The reason for both the excitement and lack of consensus is that cloud is fundamentally a new model of computing, only the third such model in the history of the IT industry, - centralized and client-server being the two previous ones. There is no single dimension around which to define a computing model, which accounts for the variety of opinions. It’s like the fable of the blind men and the elephant. Each one touches a different part of the elephant. They then compare notes on what they felt, and learn that they are in complete disagreement.
Cloud computing has significantly advanced since 2008. People are no longer starting their sessions by saying let’s define cloud computing, as this article on a 2013 cloud conference pointed out. “This is a clear indication that the industry has moved beyond elementary knowledge-gathering and onto the practicalities associated with cloud implementation and rollout.” But, different people will still have different views on cloud, depending on the particular part of computing they are touching.
I’ve been particularly interested on the impact of cloud on services, not surprisingly given my interests in service science, - the application of science, technology and innovation to the world of services. Cloud, in my opinion, is essentially the Internet of Services, providing apps, content, management, and much more to billions of people and trillions of things. You couldn’t possibly support those huge volumes of services and devices with custom, ad-hoc architectures. Cloud-based services require a much more standardized, mass customized, process oriented, industrialized approach to production, including the application of advanced technologies and rigorous science, engineering and management methodologies.
What do we mean by the production of services? Production is defined as “a process of combining various material inputs and immaterial inputs (plans, know-how) in order to make something for consumption (the output). It is the act of creating output, a good or service which has value and contributes to the utility of individuals.”
In the industrial economy, production was mostly applied to physical goods, while services were primarily labor-based. While manufacturing achieved major productivity improvements through production advances, services were considered an economic sinkhole.
This is now changing in the digital economy, as IT-based tools are bringing major technology- and organizationally-driven productivity increases to services. A fundamental transformation in services is underway. In 2009, for example, the UK’s Royal Society published the findings of a major study - Hidden Wealth: the contributions of science to service sector innovation, - which highlights the key role of science, technology, engineering and mathematics (STEM) on the evolution of services.
A recent paper, Escape from the Commodity Trap by UC Berkeley professor John Zysman, nicely explains the role of digital technologies in transforming the production of services. Zysman is professor of political science at UC Berkeley and co-director of the Berkeley Roundtable on the International Economy. The first part of the paper deals with the transformation of production, while the second, - which I’ll discuss in a separate posting, - examines the economic and policy implications.
“The emerging transformation of the production of goods and services is dramatically altering what is produced, where, how, and who captures the value,” he writes in the paper’s abstract. Production includes both manufacturing and ICT-enabled services. Their distinction is blurring as companies are increasingly looking for new business opportunities by wrapping services around their products, designing all kinds of products as-a-service, and improving the efficientcy of services by treating them as well-engineered products.
“The application of rule-based ICT tools to service activities transforms the services component of the economy, altering how activities are conducted and how value is created. We call this the Algorithmic Revolution. The development and operation of digital service applications, whether games or music services, become sources of innovation and employment. Embedding services in existing products makes the manufacture of the goods themselves a necessary, though not sufficient, basis for competitive position. This shifts the location of value creation and high wage employment.”
“In the Algorithmic Revolution, tasks underlying services can be transformed into formal, codifiable processes with clearly defined rules for their execution. When activities are formalized and codified, they become computable. Processes with clearly defined rules for their execution can be unbundled, recombined, and automated. The inexorable rise in computational power and the development of sensor technology means that computable algorithms can express an ever-greater range of activities, and a growing array of service activities are reorganized and automated. The essential point is that the codification of service activities allows the rapid replication, analysis, reconfiguration, customization, and creation of new services. It allows for business models to become more productive through extension with ICT tools and for entirely new business models to be created, offering services previously impossible at any price. The Algorithmic Revolution in services profoundly changes how firms add value.”
In the paper, Zysman discusses the wide spectrum of service activities. At one end are those activities that can be fully automated by ICT, such as ATMs in banking, Internet travel agencies and other self-service web sites, Google search, Skype and Netflix. At the other end are what he calls irreducible service activities that can only be provided by humans either because they require social skills and judgement that only human can offer, e.g., judges, psychologists, consultants; or for simple reasons of cost and practicality, e.g., hairdressers, gardeners, chefs.
In between these two poles are a wide variety of hybrid services, which combine humans and IT-based tools, such as in accounting, health care, and many different kinds of design activities. Computer Aided Design (CAD) systems, for example, enable engineers to design significantly more complex objects that they otherwise would have been able to. A variety of tools, from electronic medical records to advanced diagnostic systems like IBM’s Watson, promise to significantly improve the quality and productivity of healthcare practitioners.
The majority of services activities in the economy are still irreducible, that is, they are performed by humans with little or no technology assistance. But, given the growing power and sophistication of IT, digital technologies are being increasingly applied to activities requiring intelligence and cognitive capabilities that not long ago were viewed as the exclusive domain of humans. “[T]he deepest and broadest economic transformations are those which interweave ICT-networked, sensor-enabled products - such as nursing aids, cranes, or cars - with human delivery and judgment. The value of hybrid services depends on having human capabilities augmented by increasingly sophisticated ICT systems.”
Apple’s iPod and Amazon’s Kindle are examples of such hybrid products, where the physical devices are essentially portals to a plethora of music and books, - out there in the cloud. So are the various Internet of Things solutions like IBM’s Smarter Planet, Cisco’s Internet of Everything, and GE’s Industrial Internet.
Cloud computing is accelerating this ICT-enabled transformation of production. To understand why, Zysman explains the distinction between cloud as “where computing takes place,” and cloud as architecture, “a change in how computing is organized. . . and how the new architectural concepts are put to work.”
Cloud has significantly lowered the barriers to access computation intensive applications, making them easily and inexpensively available as a dynamic utility to be used and paid for as needed. Perhaps more important, the users don’t have to know how the underlying IT systems are configured and managed. Highly complex and sophisticated applications, - e.g., navigation, search, - are mostly handled in the cloud, without the users having to understand or deal with the IT infrastructure. In addition, cloud is ushering the badly needed industrialization of the service production factories, i.e., the data centers and overall IT infrastructures.
“A cloud architecture designed with abundant computing resources allows the infrastructure, the development tools, and the applications to be abstracted, writes Zysman. “The abstraction separates the computing stack into composable parts. Applications and infrastructure are more loosely coupled. Tightly and rigidly arranged systems, systems in which applications that create value are tightly linked to the infrastructure on which they run, are difficult to adjust and adapt. The abstraction, the decoupling both creates flexibility in how resources can be deployed and makes many of the pieces of the system into lower cost commodities.”
There are important economic and policy implications to this ICT-based transformation of production, the subject of Part II of Escape from the Commodity Trap: Will the Production Transformation Sustain Productivity, Growth and Jobs?, which I’ll discuss in a separate entry.