Exploring the path towards a unified service model
The Cloud movement is discussing the term Everything as a Service (EaaS or XaaS). In principle this is a welcome development, encouraging business and IT participants to adopt services and service oriented concepts everywhere. However it appears that the E/XaaS initiative may be more about marketing than reality. In this article we suggest how this very promising idea might be developed to clarify Cloud Service taxonomy and deliver convergence of business and IT perspectives in a Unified Service Model.
By David Sprott
Cloud computing is driving discontinuity that introduces exciting opportunities and costly challenges. Organizations need to understand these changes and develop realistic cloud sourcing strategies . . .
Gartner Group 2011
We all understand Cloud Computing is driving great change in infrastructure deployment as well as IT and business systems architecture. One of the challenges is to elevate the communication of Cloud capabilities to a business level and to integrate the concepts with the emerging Service Based Economy and Enterprise.
The concept of “Everything as a Service” has been swirling around in the Cloud community for a few months. The idea is that in addition to infrastructure, platform and software as a service there’s a whole range of service layers such as Backup, Communications, Desktop, Database, Hardware, Identity, Knowledge, Storage and many more. This is rather unfortunate as the additions all look like subtypes of the existing layers. More fundamentally the scope of “everything” seems to be constrained by Cloud technology.
To date the de facto Cloud standards body NIST has apparently not moved to embrace these ideas, and although HP in particular has been active in promoting them, it may be we need to encourage a broader definition of EaaS so that it can be more than Cloud marketing speak.
NIST has done a great job in defining general standards for the Cloud. However the treatment of the service concept in Cloud standards is less than clear. The core term service is heavily overloaded and used loosely in the Cloud reference architecture. Regardless of Cloud, there is overlap with SOA terminology and a requirement to clarify the service role as it relates to Software as a Service (SaaS), Service Oriented Architecture (SOA) and IT Service Management (ITSM and ITIL).
What’s needed is an integrated Service model that establishes conceptual convergence between business and IT, and addresses the holistic business service comprising People, Product, Process and Technology. If we then apply these generic service concepts to Cloud and non Cloud services, we are more likely to be able to address the question of Everything as a Service.
In this article we explore the opportunity for a unified service model. We look at existing service models and suggest how these can form the basis for a common core model that spans business service design, software service design and Cloud implementation and deployment.
We have been moving down the services track for over a decade. It was in the year 2000 when IBM and Microsoft led an industry collaboration to deliver the Web services standards and by so doing kick started the service and SOA market. And for the past decade Web services and SOA have remained at the top of the technology toy box. It’s only in the last couple of years that new technology toys have emerged to supersede SOA which has now morphed to become business as usual.
During that same period the entire business world has embraced services. Since the 1970s, authors like Alvin Toffler , Daniel Bell  and John Naisbitt  have predicted the post-industrial society. They forecast the end of the industrial era and the dominance of services and information. This is not a new message, the entire service provider industry has formed around this idea, and in the USA today non-manufacturing industries account for almost 90 percent of the economy. Virtually every product now has a service component to it and many products have been transformed into services.
But strangely SOA has been a highly controversial trend. The primary issue has been the fundamental tension between delivery of strategic service architecture and delivery of solutions supporting business process change. This conflict of priorities has frequently resulted in tactical, solution centered service architecture which inevitably leads to service anarchy and legacy services.
Three years ago one industry analyst went so far as to declare “SOA is dead!” Whilst this assertion was widely disputed it highlighted the issue. Since then SOA has been adopted by many enterprises, but it must be said, primarily as a technology led initiative to improve the structure of IT assets and resources. Business involvement in software services has been very low.
Perhaps as a result, genuine convergence of software services and business services has been relatively rare to date. The retailer Amazon has been a trail blazer, famously establishing the policy that all software services should be capable of being “externalized”, that is directly usable by customers and suppliers as Web services. And the result is obvious to everyone, as the Amazon business model is increasingly becoming a services platform including provision of Cloud services. The epitome of the service based platform is probably the Kindle service that illustrates perfectly how Amazon has turned into a service provider – creating a tight but highly effective relationship with its customers. We can see exactly the same pattern at work with Apple and its customers.
It’s not surprising that Web based enterprises should lead this integration of business and technology. The platform integration makes that straightforward. But more conventional enterprises are also seeing opportunities to follow this lead. Retailers are offering multi-channel, online and store based sales and service to retail consumers. Logistics companies are providing customer access to delivery services via software services. Airlines are providing multi-channel processes for ticketing, boarding card and other related services that help travelers reduce the stress of travel.
Cloud services are becoming a major priority for many enterprises. In many ways the Cloud is a service based environment supporting:
Most Cloud infrastructure environments (IaaS) also adopt SOA principles. Infrastructure capabilities are provided as invokable software services under contract. The service architecture is manifestly key to the commoditization of the infrastructure layer enabling self service provisioning and automated management.
However at other levels of the Cloud reference architecture service usage is not so formal. Whilst SaaS is being widely adopted, in my own experience much of the usage is narrowly targeted at specific business problems, and acquisition of SaaS solutions is frequently undertaken by line of business managers independent of their IT organization. Whilst this may indicate failure of IT organizations to provide adequate service, the fact is SaaS solutions’ core capabilities are not always SOA enabled, and the limitations of the architecture may not be apparent until the solution needs to be integrated into a wider enterprise application/service portfolio.
So whilst Cloud environments are inherently service based, and there is evidently widespread agreement that all Cloud architecture should be compliant with SOA principles, all Cloud layers are not necessarily SOA based. This prompts some questions:
Before we attempt to answer these questions, we need to take a broader view of what’s happening in the service world.
Figure 1 suggests that software and business services will converge as the software components of the business service become ever more central to the customer delivery. If we reflect on the transformation of Apple’s fortunes as they integrated world beating products and services, we might conclude this convergence seems highly likely to be the engine of the service based economy.
That the world is transforming to a service based economy is not in doubt. In his blog , Dave Gray argues eloquently that there is fundamental economic restructuring underway driven by the worldwide economic crisis and facilitated by the digital infrastructure and social networks.
The old world was an industrialized, producer economy, in which very large corporations became larger and larger by leveraging economies of scale. Products were commoditized; processes and systems were standardized and transactional in nature. In contrast, the new world is already predominantly a service based economy where companies create relationships with their customers and deliver value added services that are tailored to their customers’ needs. We must plan for processes to become adaptive and responsive, frequently using analytics to deliver customer satisfaction.
Figure 1: The New Service Framework
Dave Gray is stating the obvious when he says in the post-industrial economy there is an abundance of information, along with networks and mobile devices for moving that information around and the traditional information containers like documents and images, will be complemented with “things”.
The things become service platforms. Your smart phone, car, fridge, house, apartment or power supply meter all provide and consume services and are able to interact with other smart things to provide customized services and capabilities.
Of course services are not always software enabled services. There will always be physical services such as transport and logistics, health services, construction services, people services, maintenance services, the list is endless. However what we should anticipate is that most of these services will be complemented with, better enabled and potentially differentiated through software services. Which suggests service standards need to evolve; and business and service modeling languages need to converge.
There are sadly a wide variety of standards and languages in the service area. There have also been various attempts  to rationalize these disparate efforts, with minimal success. Box 1 below shows a selection of service definitions from authoritative sources. We can see the technology centric definitions mostly align around services delivering capabilities under some form of contract (or description). In principle most technology centric service models have embraced the idea that the Service is a concept that is independent of technology. They work equally well if the service is being provided by a software service or one or more physical resources that comprise a capability. However these models are intrinsically scoped from a technology perspective and, to a greater or lesser extent, lack a business perspective. They are after all, technology models. In developing them there was no remit to address a wider audience.
Box 1: Some Service Definitions
Not surprisingly the business service standards focus on people, product and process. But neither the technology or business oriented service models reflect a composite perspective of the wide array of capabilities that may be assembled to form a customer facing service. Whilst they may accommodate networks of related services, or indeed hierarchies of different types of service, they do not provide the rich model that can coordinate the multiple stakeholder perspectives, nor provide the taxonomy that makes explicit the various types of service.
There is a further question of whether the service concept embodied in both classes of model adequately represents the type of service that we will provide in future. As discussed above, there is widely held opinion that we are moving from a product based, industrial economy to a service economy in which the focus will move away from mass market transactions to individualized relationships. Existing service models are therefore predicated on a transactional world, whereas the service of the future will need to support individualized relationships.
Products are designed to be consistent and uniform with managed differentiation, for example options on an automobile allow customization, but they will always be limited to rule based product combinations. In contrast effective services will incorporate products but be co-created with customers. A business service combines people, products, process and technology resources in a customer specific version of a generic process. In a business to business situation “customer onboarding” is industry standard terminology for the customization process. In a Web based business context, the online customer is usually able to set preferences relevant to their service and increasingly interact and receive the service through multiple, concurrent channels that can be varied on a dynamic basis.
In 2004 Chris Anderson  popularized the concept of the long tail in Wired Magazine. He argued that many of our assumptions about popular taste are actually artifacts of poor supply-and-demand matching - a market response to inefficient distribution which directly reduces catalog content to optimize costs. It’s the physical, industrialized economy that puts dramatic limitations on our choice by focusing on commodities. But in the online, virtual world these limits disappear, and our ability to market niche products and services alters dramatically. The opportunity and challenge therefore for all businesses is to forge a personal service relationship with each consumer (bearing in mind this applies equally to all service markets, not just retail).
Customization is always a major challenge in all service deliveries. There is a natural tension between the provider who is interested in standardizing service elements and reducing onboarding costs where possible, and the consumer who wishes to minimize integration costs and integrate with existing technology and processes. Into this model we also need to throw into the mix the concept of structured and unstructured behaviors.
This is particularly relevant to physical aspects of complex service deliveries, such as requirements to substitute resources with varying SLAs, or introduce temporary rules to accommodate highly unusual circumstances for reasons outside the provider or consumer’s control. Whilst we might like to think we can design highly structured online services that cover the totality of required behaviors, there is an implicit competitive challenge to respond to unstructured situations to improve the quality of the delivered service in some dimension such as security or safety. And if this topic seems a little esoteric, consider the services offered by or consumed by a robotic device  which operates in a hostile environment and is has delegated responsibility for its actions.
This discourse suggests a unified service model needs to articulate more than the basics of the service, provider consumer relationship. It must encompass a richer set of characteristics that reflect tomorrow’s service based business.
These could include:
From the foregoing it will hopefully be clear that we need more than simply a great debate about definitions of the service concept and a merge of the various service models that already exist. We need a richer set of models and an appropriate level of convergence or alignment together with a rationalized taxonomy that clarifies type of service.
Continued in PDF