The word service is one of those general terms that have many different meanings depending on the context. In different situations, service is what you get in a restaurant or gas station, what you're in when you join the army, or what you endure in church. Web Services, of course, have come to mean something more specific, namely software services that consist of standards-based interfaces to IT functionality that one computer system exposes to another. What, then, is a "software service"? We can define a software service as any functionality that one computer provides to another, but this meaning is quite broad. After all, we could be talking about AOL or email, and we'd still be covered under this loose definition of service.
It makes sense, therefore, to delineate the various types of software services, so that we can all make sure we're speaking the same language. In this ZapFlash we're going to discuss shades of meaning for the term "software service," not simply as an exercise, but because there is a convergence taking place in the marketplace today among the various types of software services. This convergence promises to be disruptive in many ways, and will flesh out the principles of Service Orientation beyond the still-emerging world of Service-Oriented Architecture (SOA).
Four Kinds of Software Service
First, let's peel back the onion of software services a bit, and clarify the different types of services:
- Software-as-a-service -- IT capabilities that customers can access as hosted software functionality via thin client interfaces, rather than by installing applications on their local machines or networks. The concept of software-as-a-service combines the technology model for accessing distributed functionality with a business model that includes charging for that functionality. Good examples of software-as-a-service are Salesforce.com or the application service provider (ASP) business model that first emerged in the late 1990s.
- Customer-facing services -- capabilities vendors provide to customers on an ongoing basis, sometimes with a corresponding ongoing revenue stream. Such customer-facing services include free and premium online services such as Yahoo! Mail or AOL.
- IT services -- capabilities that an IT organization provides to its users. Examples of IT services include corporate email and print services.
- Loosely coupled Services -- contracted interfaces to software functionality and data that independent software consumers can access. The loose coupling depends upon a software architecture oriented toward such Services, in other words, SOA. These are the Services -- often Web Services, but not necessarily -- that ZapThink focuses on. We always capitalize this meaning of the word "Service" to distinguish them from all the other meanings of the word "service."
Areas of Service Convergence
You're probably already thinking that there's some overlap among the types of service we described above. After all, we could easily consider Yahoo! Mail to be software-as-a-service, and some IT services (like portals, for example) might also fall into that category. When loosely coupled Services get stirred into the mix, however, the story focuses more on convergence than on the various meanings of a general term. Here are some examples of how loosely coupled Services lead to the convergence of the different types of service:
- Software-as-a-service offered as loosely coupled Web Services -- instead of requiring thin client interfaces, a wide range of different types of consumers can now access software functionality offered over the Internet as remotely hosted, standards-based, loosely coupled Services, including emerging Rich Internet Application (RIA) solutions, desktop software applications such as Microsoft Office, and automated business process flows. When the provider of such Services leverages SOA to provide loose coupling, it can offer multiple versions of Services by exposing different Service contracts, and even route Service requests to the appropriate Service at runtime.
- "Headless," discoverable IT services -- typical IT services have rigid, proprietary interfaces. For example, companies typically implement print services by configuring computers with the appropriate software, and then provision users to access the appropriate printers. If printers exposed loosely coupled Service interfaces to any interested party on the corporate network, however, then any authorized service consumer could go out on the network and look for them, learn about them, and access them. They would then be "headless" in the sense that the print service wouldn't need its own user interface. Instead, the consumer would provide whatever UI was appropriate. Such print services were part of the original vision behind Jini, by the way -- only now, this vision doesn't require a particular language or runtime environment.
- The rise of the Service marketplace -- originally, AOL's offering consisted entirely of rich content via a proprietary interface. Next, they added a Web interface with some Web-based content. Now, what if AOL or MSN featured Web Services as simply as they currently offer rich content? Furthermore, what if any desktop application you were running were able to consume relevant Services from such an online customer-facing service? For example, AOL could provide content via a Service that you might consume with a spreadsheet or other application. Just as the Internet and World Wide Web enabled the broad distribution of both free content as well as innumerable business models based on premium content, Service convergence will lead to an explosion of loosely coupled Services that customers can consume on a free, licensed, subscription, or per-transaction basis. The more that companies expose Services for customer use, the more that companies will want to create such Services. Such a market is already growing, and approaching what we call the Web Services tipping point as available Services increasingly compete with one another for business.
Consequences of Service Convergence
As SOA enables the convergence of software-as-a-service business models, customer-facing services, and IT services into broadly applicable, loosely coupled Services, the disruptive nature of this convergence will lead to many changes, both within IT as well as to the businesses that leverage such Services. Here are some of the consequences that Service convergence will have within IT in particular:
- Convergence of SOA management and other forms of management -- in one corner we have IT service management (ITSM) and business service management (BSM), and in the other corner we have SOA management. ITSM is now reasonably mature, as companies currently have the tools and best practices to effectively manage IT services. BSM is less mature, partly because both enterprises and vendors still struggle with aligning business needs with IT capabilities. Meanwhile, SOA management, the logical evolution of Web Services management, helps companies manage their SOA implementations. As the different types of services converge, however, SOA management will subsume both ITSM and BSM, breathing new life into BSM by more closely tying business needs with IT resources.
- The evolution of the traditional desktop application -- taking the combination of RIAs and the converged world of Services to the next logical step, the days of a desktop application as an isolated, unchanging, and discrete piece of software are numbered. It will no longer matter which parts of the application are local, and which are remote via software-as-a-service. Whether pieces of the app are browser-based or browserless RIAs will also become irrelevant, and even the features shipped with the product on day one will simply be the starting point for what the product can offer the user over time. Users will essentially be able to assemble applications as they require out of abstracted Services that may either be local or remote. So, welcome to the new, Service-oriented desktop application of the future!
The ZapThink Take
There is already substantial buzz in the marketplace about the convergence of the various kinds of service into a single, SOA-supported ecosystem of software-based Services. This convergence promises to be disruptive, both within IT as well as in the business world. If the companies that choose to build their businesses on converged Services get this convergence right -- in particular, by basing it on SOA -- then they will truly have a bull by its horns. If they drop the ball on the underlying architecture, however, then the benefits of service convergence detailed above will still be largely out of reach.
Unfortunately, there is a substantial danger that this convergence of service types might not happen at all. After all, leveraging the power of SOA to achieve such convergence is disruptive in the sense that businesses will themselves transform, and many will fail simply trying to get their own architecture house in order. Therefore, such convergence represents already signs of advancement for companies deploying SOA solutions with and outside their organizations. As Service convergence takes place, the real action will be how this convergence impacts relationships between companies, and between organizations and consumers. The buzz will likely not center on SOA specifically, but SOA will necessarily be in the background making the convergence work. In order for companies to realize value that's more than simply the next version of AOL or MSN with a smattering of disconnected services or the next version of a desktop-installed enterprise print service app, companies must leverage the power that SOA offers in order to make true service convergence, in all its definitions of the word, a reality.
This was first published in December 2005