- The Information Age and the Internet
- The Web of Services
- The Ecosystem
This chapter is from the book
This chapter is from the book
This chapter is from the book
1.2 The Web of Services
In today's Web-centric world, it is convenient to think of entities or resources that are connected to each other through some mechanism — by means of a hyperlink in most simple form. The primary purpose of the WWW is to make it easy to share information among interested parties. The pages, together with the hyperlinks, create a web of information. Any visitor to a page can jump to another by following the hyperlinks or by explicitly providing the Uniform Resource Locator (URL) for the page of interest. This is the state of most of the World Wide Web (WWW) as we know it today. This infrastructure is what also facilitates the e-commerce applications today for companies that want to reach their customers over the Internet.
We can extend the concept of the Web to include not only pages, but more active components of the Internet — namely, the services. A web of services is similar to the web of pages in a lot of ways. For example, they will be connected to each other through some mechanism, and their users will be able to go from one service to the other using these links. However, the service web that e-Speak envisions is also very different from the web of pages. It is these differences that distinguish the web of services from the web of pages. The principle differences are in the form of service interactions. We can classify these interactions in three different categories:
User-to-Service interaction
Service-to-Service interaction
Service Deployer-to-Service Deployer interaction
1.2.1 User-to-Service Interaction
We have been talking about the services until now without any specific definition for that word. Although we all have some notion of what a service is, a formal definition will be useful for the rest of the book. It must be noted that the definition we give below is not expected to be something that will be all-encompassing but rather a working description of the notion of service that will be used in this book.
Service: A service is an active program or a software component in a given environment that provides and manages access to a resource that is essential for the function of other entities in the environment.
Figure 1.2 depicts a service. It is important to note that the notion of resource is rather broad in this definition. A resource could be a piece of hardware, such as a hard disk, or software, such as a math library. It could also be some data in a database or some information, such as a news item. This definition is somewhat simplistic yet very powerful. It is simplistic because it tries to be as broad as possible and specifies very few specific characteristics or restrictions on qualifying entities. A service need only be a program or a library function to qualify. The key, however, is in provision and management of a resource. Provision means that it has to make the resource available to other entities, and management denotes controlling and granting the access rights. You can see that this automatically mandates interaction with other entities in its environment. Thus, a program or a software component that does not interact with other entities will not qualify as a service. An equally important aspect is controlling a resource that is desirable by other entities in the service environment. Any resource that is useful only internally to the program does not help in making that program a service.
)
Figure 1.2 Resource access via a service.
As can be noted from this definition, a service is an active component, whereas a Web page is a static, one-time representation of some information. One can interact with a service (to access a certain resource that the service manages), and the service responds back with a response (rejecting or providing access to the resource). One could argue that a page can do something similar but, in fact, it is the Common Gateway Interface (CGI) program or the servlet behind the page that will be responding to the request. The page provides the means for communication between the user and the program or the servlet. It is the CGI program or the servlet that is acting as a service in this case.
The interactions with active components as services are richer, compared with the passive components, such as Web pages. A service can show a high level of dynamism while catering to the user. Due to this a high degree of customization, flexibility can be achieved, thus giving a better user experience. So far, we have been delivering the services through a series of Web pages. So how is this dynamism exhibited today?
Most of the initial Web content was on static Web pages. When the CGI technology was invented, it became possible to connect programs to the Web pages. These programs could be used to process user input (such as a login or an order status information request) and render a complete Web page dynamically. This dynamically generated page would contain information specific to the user request. With the advent of style sheets and Java/Active Server Pages (JSP and ASP) technologies, most of the Web pages can be expressed in terms of the overall style, with the connected program filling in the specifics.
It can be noticed that there are a lot of standards on the presentation aspect — the browsers, presentation language Hypertext Markup Language (HTML) and style sheets. However, we are missing on the standards on the actual interactions these presentation technologies are supposed to facilitate. How should access to a resource be provided? How should the user and the service agree on describing a resource and describing the usage of that resource?
In a service-centric world, the emphasis is put on answering such questions. The presentation is a final-stage consideration when it is time to deliver the resource to the user on a specific medium. In a web of services, the user is thus free to choose the appropriate medium of delivery at a specific time. This, coupled with presentation-level flexibility, can produce a highly customized user experience.
The emphasis on the description of various entities also opens up quite interesting possibilities in discovering services for a user or creating a new business partnership to enhance a service offering of a business. As long as the user and a collection of similar services agree on the terms of description, the user can, in theory, choose any service to accomplish a task. The choice could be based on a variety of criteria such as past experience, third-party rating, or even user's whim at that moment. Regardless of which service is chosen, it is possible for the users to express their tasks in a consistent, previously agreed-on manner. From a cognizant user's perspective, this means that there is now a choice about which service to choose based on certain criteria — the description of the service class. Such a user can choose any service after due diligence. On the other hand, from an non-conscious user's standpoint, it doesn't matter which service serves your request so long as you can describe your needs well and the service is capable of understanding your description and acting on it to fulfill your request. Imagine buying a car. One can do all the necessary research about car dealers, provide the description of the car, float a request for quote to all the "qualified" dealers, and choose the dealer that has the best price. On the other hand, one can also just find the nearest dealer and buy the car so long as the price sounds reasonable. Between the two scenarios, the users did not agree on the best way to choose a service, but in either case, they could reach a service that could fulfill their need. There is only one fundamental thing the users and the dealers needed to agree on — the description of a car. There is one more aspect to the user-to-service interaction —discovery. We already discussed the description for a class of services. Discovery of a service means choosing, either deliberately or accidentally, a specific service from a class of services, using a specific description. Intelligent matchmaking like this can ease the pain in finding and interacting with the right service.
To extend the idea further, imagine now a task that is so complex that no single service can do it — such as the trip-planning task we described earlier. In such a case, to accomplish the task, we will have to choose a set of right services and interact with them individually. As we will see in the next section, a service-to-service interaction mechanism can help solve this problem effectively. Such a composition of services will create a seamless experience for the user.
In the e-Speak vision of the services, the user-to-service interaction is thus characterized by:
Service description
Service discovery
A high degree of customized experience
Service composition to accomplish higher level tasks from services acting as basic building blocks
You will notice that e-Speak's vision describes a special class of services. The services from this class are highly interoperable in offering their services. Such services are called e-services.
E-service: An e-service is a service that abides by a specific framework to offer its services. The framework provides the means to describe and discover the service, audit its service offering, and integrate the e-service with other e-services to offer higher level e-services.
The industry has adopted the term web services for this class of services. This book concentrates on the service-centric computing, including both e-services and web services. As a result, both the terms web service and e-service point to this class of services for our discussion and we have used these terms interchangeably in the rest of the book except in Chapter 13.
1.2.2 Service-to-Service Interaction
A user of a service can expect a significantly different experience while interacting with it, as we saw earlier. We also mentioned that the e-services are highly interactive components and communicate with other e-services. Such interactions could be either to use the service offered by an e-service or to collaborate together to form a higher level service that can be used to add complementary or additional value. In either case, the interacting services should have an experience at least similar in nature to the service-to-user interactions we described previously. Should the interaction be richer than that?
Let's look at the service-to-service interactions more closely. By the definition, a service is an active software component. Thus, the service-to-service interaction inherently implies at least two software objects communicating with each other. Such an interaction is different from either the Web page interaction or the service-to-user interaction. Compared with a set of Web pages joined together by hyperlinks, a service-to-service interaction is very dynamic in nature. In the Web page realm, the hyperlinks (or the connections) are either purely static or (maybe) somewhat dynamic. In the service world, the connections formed are determined and bound on the fly, based on the requirements. Take the trip-planning example from the previous section. In that case, a trip-planning service can decide which airline reservation service to collaborate with dynamically, based on the user's preferences, which could include using a specific airline, lowest price, shortest travel time, or the destination itself. Note that this search can result in finding airlines or airline reservation services that the trip-planning service may not be aware of until that moment! It is this level of dynamism that differentiates a service-to-service interaction from the usual "Web page"-like interaction. E-Speak is designed to support such dynamism.
The other aspect of service-to-service interaction is the notion of dialog between the two interacting services. A service can have a dialog, a bidirectional communication, with other services. In the passive Web page world the absence of such a dialog is near total. A Web page can redirect you to another Web page, either automatically or via a hyperlink without giving any prior indication to the other page. Also very rarely is there any communication between two pages.
NOTE
There are a few examples of unidirectional communications between Web pages. A Web page, while redirecting the user to the other Web page, can communicate information regarding itself to the Web server or the Web page being loaded. The server or the page can use this information to present the page information differently or to maintain a log. Various business models have been built around such a capability. One such business model is revenue sharing. When a Web portal redirects the user (automatically or by conscious user choice) to another company's Web page selling a product, the Web portal is compensated monetarily for the redirection. A product company may have such an arrangement with several portals, and vice versa. In such a case, for each redirection, a context for redirecting the page must be communicated to the product company's Web server or Web page. This context is then used to determine the revenue-sharing amount for each portal and the participating product companies.
The possibility of bidirectional communication can help in creating a more customized and richer set of interactions. As we will see later in this book, such interactions make it possible to introspect a service to understand its description, information-sharing format, and workflow sequences. Based on this knowledge, one can qualify (or disqualify) a service or amend its own workflow to accommodate (or abide by) the other service's workflow. Again, this is an aspect of the dynamic behavior that is absent in the web of pages. It should be noted that such a level of dynamism is also not typically exhibited in the service-to-user interaction. A user usually has a specific behavioral pattern, and this pattern is difficult to change to accommodate a found service. In fact, in most cases, systems are designed to accommodate this behavior pattern. In the e-Speak world, as well, the same principle follows. It is summarized in the e-Speak motto: To make the Web work for you and not make you work on the Web.
1.2.3 Service Deployer-to-Service Deployer Interaction
A rich interaction pattern between a service and its users or other services can create an environment that can also influence the interaction between the service deployers. In the current state of the Web, the relationship between various service deployers is rather static — at least in the electronic world. Service deployers enter into an alliance to connect each other's services. The agreement to form this alliance happens in the paper world or from the electronic world's perspective —offline.
After an alliance is formed, a team of engineers gets to work to build the electronic bridges between the two companies forming the alliance. The electronic bridges built are usually in the form of a Virtual Private Network (VPN) or a collection of Web pages secured through some mechanism, such as digital certificates from either side. Whatever the form of these bridges is, veteran programmers can tell you that it takes some significant effort to build them, and the maintenance of them is even more cumbersome. In the meantime, the executives at each company are busy forming new alliances to expand the business or severing the ones that do not live up to the expectations. This activity between the business managers and technologists in a company creates a constant state of flux in which the electronic bridges are constantly being built or torn down — each of which is a labor-intensive endeavor.
The problem lies in the fact that the semi-dynamic technology-level solutions cannot live up to the expectations of hyper-dynamic world of business alliances. The efforts required to build or tear down electronic linkages between two or more companies comes in the way of rapidly forming lucrative alliances or pulling down un-profitable ones. Currently, the technology folks are trying hard to keep up with the business people on this issue.
If, however, it is possible, to find dynamically a collection of services suitable for the task you want to accomplish on the fly, as we discussed earlier, it should also be possible to form the alliances (or break them, as the case may be) at the same pace. The e-Speak infrastructure is designed to make this possible. Whereas e-Speak provides the description of a service to find it, it has the terms of use and other constructs to communicate the modalities of the service being offered. These e-Speak constructs, although termed a contract, are somewhat rudimentary and are not quite as powerful as a contract in the real world. However, it is possible to devise such instruments of business operation using them. It must also be noted that the legal system around contract formation in the digital world is not fully developed as yet, and without that, the effectiveness of even a sophisticated implementation will also be limited. However, the e-Speak implementation is a step in the right direction, and as we mentioned earlier, it is possible to create a more advanced mechanism for contract formation using it.
Equipped with the dynamic discovery and contract formation on the fly, one can imagine a very different world. In such an environment, it would be possible to offer a service and reach customers you didn't even know existed. Also, from a customer's point of view, it would be possible to describe what they want and find a set of services that fulfill it. The discovered and eventually selected services will vary every time, based on the parameters the customer provides, such as location, rating, and preferences. For the chosen set of services, one could retrieve the terms of use and other quality-of-service aspects and make a decision on which service to use. In fact, instead of going through the whole process, we could delegate that task to another service we trust!
As you can see from the description above, the service-oriented world like this means a large variety of service providers. Each of the service providers by themselves are providing only a small piece of functionality — a functionality that is their core competency. However, when they are together, connected through the mechanisms described by e-Speak, they can create a very powerful service-providing environment. E-Speak calls such an environment an ecosystem.