Web Services and Internet 3.0
SOAP exists because of an idea called Web services. Web services are new, and there aren't a whole lot of them out there right now. However, they're growing in number daily and have the same potential for explosion that Web pages had back in 1996 (maybe even greater). To get a strong idea of what Web services are, and why they're part of the next phase of the Internet, it's useful to cover a bit of Internet history to get some context.
A Brief History of the Internet
The Internet began as an American military experiment called ARPANET, whose goal was to build an interconnected computer network that could continue to function even if some nodes (that is, some computers) were destroyed or rendered inoperable by, say, a nuclear strike. Clearly, this experiment worked (without having to test exactly what effect a nuclear explosion would have), and ARPANET grew until it became known as the Internet. Most serious geeks knew about it and started playing with something called email, even though everyone at the time knew that phones were more convenient. Some folks have called this Internet 1.0.
Then in 1991, a guy called Tim Berners-Lee came up with a way for Swiss physicists to post their scientific papers on this Internet so they could read each other's papers easily. This idea caught on quickly and bloomed into the World Wide Web. Initially, the Web was used for people to communicate with each other, sharing such things as scientific papers, personal Web sites, marketing sites for companies, and pornography. In the beginning, the Web was mainly a static environment used mostly for person-to-person communication.
As more serious developers started learning about the Web and a new technology called Common Gateway Interface (CGI), they were able to create programs that allowed people to actually accomplish some things over the Web, such as buying books, reserving airline tickets, and transferring money across bank accounts. In this fashion, a large portion of the Web became devoted to people-machine communication and interaction. People interacted with software programs in order to accomplish things. This phase has been called Internet 2.0.
Many people think the next step in using the Internet will involve software applications communicating with each other over the Internet. These programs will use each other's components, modules, and methods in order to accomplish their tasks. For example, you could write a program in Java that does something, but your program needs to get some information that lives on another machine, a remote computer. Your Java app must be able to communicate with this remote machine in order to get its work done. It must be able to ask that computer for the data it needs, and be able to receive and understand that data. The remote machine must be able to listen for your Java app's cry for help, process the request, and return the information in such a way that your Java program can actually read it.
This next use of the Internet that focuses on machine-machine communication has been called Internet 3.0.
So What's a Web Service?
A Web service is the name for a method or function that is available for other applications to access over the Internet. To use a classic example, say that you have a program that needs to find out the latest stock price for a certain company, and that information lives on another computer. This remote computer has been configured in such a way that if you send it a request for a stock price, it will respond with that company's current price. In this case, the remote computer is hosting a Web service. It's just a method or function that can be accessed by other programs over the Web.
Another way to look at it: When people want some information, they look for certain Web pages and get that information in HTML form. When applications want information, they look for certain Web services and get that information by requesting it from another computer.
People receive Web pages in the form of Hypertext Markup Language (HTML). Computers receive information from Web services in the form of SOAP. SOAP is HTML for computer-computer interaction over the Web.
The goal of Web services is a seamless integration of services across the hardware and software barriers that exist on the Internet. In this system, much of the Internet could become repackaging of other bits and pieces out there.
Money and Web Services
Why would you bother to write a Web service? The story to get developers such as yourself onto the Web services wagon is that you could write a Web service, expose it to the world, and then charge people who used your Web service, either on a per-use basis or subscription basis. It's certainly an appealing idea: Write something brilliant, even a small thing, and people will flock to you to use your program and drive truckloads of cash to your front door. This has not actually happened yet, but honestly, the potential is there. The next few years will be interesting.
Evolution from Software to Services
Right now, the idea of software is that of a discrete process. A consumer buys a shrink-wrapped package with a manual and a CD. They install the CD's program on their machine, and it runs without anyone else on the planet knowing about it. Then the person usually "loans" the CD to someone who asks nicely.
In the Web services view of the world, this process disappears. Instead, you buy access to a Web service for a certain amount of time. For example, instead of buying Word at the store, you'll buy the access rights from Microsoft for, say, a year. You download some components from the Microsoft Web site, but otherwise, your Word program mostly lives on some server at Microsoft. If Word needs to be updated, Microsoft simply upgrades the program on their server, and your program is updated without your even being aware of it. This upgrade is free for youit's part of the subscription price. After a year, you must renew your subscription.
Notice you don't have a CD to loan to anyone else. Part of the push for Web services is that they will cut down substantially on software piracy (at least for a while, anyway). With less piracy, more people will actually have to buy software, and software companies will make more money.
Though this probably rankles some of you, this has aspects of fairness to ityou pay for what you use. Most of us more or less adhere to the philosophy of using something for free if you're using it to learn, but paying for it once you're using the software in a way that makes money for you.
And this is one big reason why Microsoft is putting so much of their weight behind Web services and making it easy for developers (and themselves) to create Web services: In the long run, they think they will make more money.
Requirements for Web Services
Initially, a lot of people will be experimenting with creating their own Web services, and many of those services will break from time to time. However, if you want your Web service to be used by anyone who will actually pay money for it, there are some things you must remember. Your Web service:
Must be 100% reliable. It must always work, and always work in the same way.
Can also use other Web services, which means that everything must be working or a whole chain of services could go down.
Must be extremely availablethat is, highly scalable and always on.
Must be able to handle unexpected inputs (usually returning a fault is adequate).
Must not affect the availability of any other Web service, even if those Web services share common components.
Must be secured so that only authorized systems and users can use it.
That's a high-level look at what will be required of Web services in order to become used and (hopefully) profitable.
From a more granular point of view, what do Web services need in order to be able to communicate with each other and announce themselves to the world? Here's a list of some other requirements. We'll go into more detail about these later (actually, the whole book is devoted to understanding how the following requirements are implemented, so don't worry if the list below isn't 100% clear; it will be).
A standard way to represent data, so the services know what to look for.
A common, extensible message format.
A common, extensible service description language.
A way to discover services located on a particular Web site.
A way to discover service providers.
Some Concerns about Web Services
Famed software interface guy Bruce "Tog" Tognazinni recently exposed a potential abuse of Web services. He signed up with a digital TV recording service (ReplayTV), and one of the reasons he signed up was because he could record TV without the commercials. He paid the subscription price and received the service. Partway into the year he signed up for, the company changed its service, and he could no longer record without commercials. Essentially, he paid for a service that was now significantly less valuable to him, and there was nothing he could do about it.
ReplayTV has since changed their policy and reverted back to their original functionality, but this raises a legitimate concern: How well do you trust the company that is providing a Web service? In an exaggerated example, suppose you're using a Microsoft Web service to use Word, and one day they decide they don't like Word anymore and you find the only thing the program does is run Tetris.
As with many other aspects of Web services, we'll have to see how this issue plays itself out. I can already picture attorneys wringing their hands in happy anticipation of conflict.
Finding Web Services
We'll cover this in more detail later, but much work has been done to define exactly how someone can locate a Web service they need, how they can find out what sort of input parameters that Web service needs, and how to structure those parameters. The current plan is for all Web services to register in a big list called the Universal Discovery, Description and Integration Registry (UDDI Registry). Inside this registry you'll find all the information you need to connect to and interact with a particular Web service. We'll cover this in more detail in Chapter 6, when we cover UDDI.
Once you find the Web service you want to use, your application will also have to know what sort of information to pass to this Web service. A whole new language called, appropriately, Web Services Description Language (WSDL) is being developed that spells out exactly what sort of information a Web service needs in order to work.
Web Services and RPC
The request-response form of communication between a local computer and a remote one is often called a Remote Procedure Call (RPC), because it involves a program invoking a method, or function, that lives on a remote computer. RPC is not newpeople have been doing it for about 20 years in one form or another. Sun Microsystems is usually given credit for being the first to create a generic way for applications to call methods on remote machines.
This method of software applications talking to each other over the Internet, machine-to-machine communication, will likely be the next big use of the Web, and may surpass all others in significance. Will it actually happen? Probablymany developers are excited by the idea, and incidentally, Microsoft has bet the company's future that it will.
This kind of system is also known as "distributed." That is, the system is distributed over several computers. It could be said that part of your application lives in a separate place (even though you can also say that your application is simply calling another application). In any case, this is known as a distributed architecture.
A view of how this works is illustrated in Figure 11.
FIGURE 11 A Web service using SOAP
In this scenario, your application wants some information that lives on a remote computer somewhere.
Your application sends out a request for data, and the first stop on its journey is the SOAP client (which might also live on your computer).
This SOAP client takes in the request for data and translates the parameters into a SOAP message, like we saw at the beginning of the chapter.
The client then sends this SOAP message on its way to the remote computer, which has a SOAP server running at all times, listening for incoming SOAP messages.
The job of this SOAP server is to listen for SOAP messages, take the input data in those SOAP messages, and translate it into something the remote method can understand.
It then passes the input data to the remote method.
The method does its thing and comes up with some data.
This data is given back to the SOAP server.
The SOAP server then acts as a SOAP client, in that it takes the returned data and turns it back into SOAP.
It then sends this response back to the original SOAP client on your computer.
Your SOAP client then translates the SOAP data into something your application can understand.
And finally, your application gets the data it needs in a form it can understand.
A vital facet of Web services is that it doesn't matter what platform the requesting or responding computer is. It could be a Mac calling a method on a Linux box, or a Palm requesting something from a Windows machine. Not only do Web services not care about platforms, they don't even know. This is like the Web: When you download a Web page, do you know what kind of server that it came from? Not without doing a little research. And really, you don't careyou just want your Web page. Web services have a similar goal and so must be platform-independent.
Web services take this concept one step furtherthey don't even care what language the remote methods are in. C# can talk to Perl, which can talk to Python, which can talk to PHP, which can talk to a really fast guy with an abacus. It doesn't matter. The only thing these programs need to have in common is their ability to read and write SOAP messages. Some of you may remember when people were trying to make a new language called Esperanto, a language that everyone on the planet would learn to speak and read so everyone would be able to talk to each other. It failed miserably. Fortunately, computers are more malleable than people, and we can force computers to all speak the same language, and that language can be SOAP.
Transparency and Plumbing
A Web service using SOAP is an example of a loosely connected application. Connections are said to be tight or loose. The looser the connection is, the less one application knows about the other. Web services are very loosemethods only know what other methods do, not what platform they're on or even what language they're written in. Loose connections generally go through an intermediate step. In the case of Web services, SOAP is this intermediate step.
The details of the connections between applications are known as plumbing. The tighter a connection, the more you, the programmer, must know about plumbing. Plumbing is usually a pain, and most developers avoid it if possible. That may be one reason why connections have become looser and looser over the years: No one wants to program plumbing.
Looser connections generally provide more flexibility for developers.
Support for Web Services
Companies like Microsoft, Oracle, IBM, Hewlett-Packard and Sun are running to create their own Web services. In fact, that's what Microsoft's HailStorm project is: a bunch of Web services.
Part of the Web services plan is for people never to install software on their computers, but instead just access the programs over the Internet. For now, "Web services" just means invoking methods over the Internet: My computer makes your computer run a little program for me and return the results.
You can create Web services right now. You don't need to wait for anyone, including Microsoft. This is important: .NET is not necessary to create a fully functional Web service. Some .NET tools (like Visual Studio.NET) are designed to make it easy to create Web services, but that's itno one, not you or anyone, has to go through Microsoft to create or use Web services.
In general, it's expected that a number of individual developers will slowly start experimenting and creating their own Web services until some sort of critical mass is reached. At that point, an explosion of Web services is expected to arise as businesses eventually decide that Web services are a solid technology to invest IT capital in.
Why SOAP Is Needed
Now, as developers became more sophisticated in both their skills and needs, they became frustrated with some parts of the Web. It's not because their dot-coms sank beneath the waves along with their dreams of IPOs, but because of the inherent limitations of HTML. Right now, the easiest way to get information from another computer is to request a Web page, and this Web page usually appears in HTML form. It's almost impossible for an application to figure out what the important pieces of data are in an HTML file. After all, HTML is mostly a bunch of formatting tags: bold, italic, table cells, and so on. Here's one way to look at it: HTML isn't information, it's just a picture of information. That's one reason why HTML works so well for people and so poorly for machines. By and large, we see in pictures, but computer programs need consistent, predictable structure in order to make sense of the world. Thus, computers need something other than HTML to talk to each other.
You may remember from earlier in the chapter that we discussed some of the requirements for a Web service. Two of them were:
A standard way to represent data, so the services know what to look for.
A common, extensible message format.
SOAP provides the answer to these two requirements. More precisely, XML provides a standard way to represent data, and SOAP provides an extensible message format ("extensible" essentially means you can make up your own tags; for example, HTML is not extensible).