Wireless and Web Services: Beyond the Hype
Wireless and Web Services: Beyond the Hype
By: Larry Mittag
Jan. 1, 2000 12:00 AM
Wireless data communications is a fascinating technology. The idea that a
clean digital signal can be extracted from the analog soup of RF
communications is audacious in concept and even more so in the
implementations that are either currently available or being planned for the
next few years. Given all of this, why is it that no one seems capable of
making any money on it?
The simple, brutal answer to this question is that customers will not
pay for raw technology, no matter how elegant it is. They will admire it and
discuss its merits and flaws in chat rooms, and they may even dabble in it.
But when it comes to using it as the cornerstone of a new business
application or integrating it into their busy home lives, there must be more
behind it than elegant technology.
There must also be solid evidence that it is going to do
useful things for them in a dependable, economical way.
To date, this evidence has largely been missing among
the smoke and mirrors that have been characterized
as the "Wireless Web." That hypefest has
played its course. Now it's time for wireless
data communications to get serious.
The simple, brutal answer to this question is that customers will not pay for raw technology, no matter how elegant it is. They will admire it and discuss its merits and flaws in chat rooms, and they may even dabble in it. But when it comes to using it as the cornerstone of a new business application or integrating it into their busy home lives, there must be more behind it than elegant technology.
There must also be solid evidence that it is going to do useful things for them in a dependable, economical way. To date, this evidence has largely been missing among the smoke and mirrors that have been characterized as the "Wireless Web." That hypefest has played its course. Now it's time for wireless data communications to get serious.
Building the Base
The amazing thing is that so far, people have been debating these technologies as if it matters. The reality is that there are a few basic criteria that must be in place for a wireless infrastructure. These minimum requirements will build the base, which will in turn support the next level of services where customers will see value in the networks.
These minimum requirements are:
These requirements are necessary, but are not, themselves, sufficient for the widespread use of wireless data. They will provide the foundation, but foundations are not really that interesting until someone builds on them.
Web Services: The Next Layer
In fact, it is also possible to do this under second-generation PCS connections. I still remember getting my e-mail on my laptop while traveling 70MPH (or maybe a little faster) in a Corvette on the road to Las Vegas (my wife was driving at the time). Then I could forgive the fact that the connect speed was an anemic 9600 baud and that the true throughput was much less than that. The novelty was enough then, but that mode is not sustainable. That was then and this is now.
Truly mobile wireless applications must move beyond this "compatibility mode." They must capitalize on the things that wireless connections do best. They must be capable of taking advantage of intermittent connectivity to pass small, self-contained and self-describing packets of data that can provide useful information. The applications must be able to request very specific information and to aggregate these packets over time rather than expecting a full major transaction to take place every time a user looks at the device. In short, these applications must be built to take advantage of Web services.
This is where the smoke and mirrors come in. Microsoft has done a masterful job of making everyone aware of Web services through their .Net advertising blitz, but as usual, there is not much detail to be had as to exactly what Web services really is. How exactly are the customer's apprehensions about black versus red for that spiffy new car in the Microsoft commercials communicated to the automotive factory floor in real time? Setting aside for the moment the question of whether or not it is a good idea to do this, the process is a good example of how Web services can be used.
The first step is the decision to buy a car. This opens a transaction, a fact that is signaled (perhaps prematurely) by an anxious salesperson on his or her portable tablet. The tablet puts together a sales request in the form of an XML record that describes the car to be bought. This record is sent to the factory server, which would be capable of sending back relatively exact information on when the car can be delivered. It could also begin actually building the car, but that might not be a really good idea since the order obviously has not yet been solidified. Luckily, in this case the salesperson in the commercial saves the day by having extraordinary bad taste in shoes, but we certainly can't depend on all salespeople being that tacky or can we?
The point is that the tablet computer in the hands of the salesperson communicates directly with the computer in command of the assembly line. This is the one-degree of separation to which the Microsoft commercial campaign refers, and it is also the promise of Web services. They are all about allowing computers to communicate effectively among themselves. There are real advantages to this in the universe of wired connections, but those advantages also extend significantly into the realm of wireless data connections.
One of the primary advantages for Web services in wireless communications is the use of self-contained XML data records. This maps much better into the intermittent connectivity of a wireless connection than a session-level transaction that requires a constant connection. Web services define a transaction as a sequence of independent exchanges of data records rather than as a monolithic exchange under a session. The flaw in the procedure shown in the Microsoft commercial is that the paint should not have been applied until the customer was sure (i.e., the transaction was finalized).
The real power of Web services is not getting the order to manufacturing quickly, though. It is more in the capability of having manufacturing interact directly with the customer. In this case the factory might have been out of red paint, so the order for a red car would have triggered a reorder of red paint. Simultaneously, the response to the order sent back to the customer could have included an exact delivery time including the time spent waiting for the paint delivery. The customer then has the choice of getting a black car tomorrow or a red one the day after. In this scenario, manufacturing is very closely connected to their customers as well as to their suppliers.
Ultimately, the principles of "Just-in-Time" manufacturing and delivery could be brought to their logical conclusion, where the order triggers the order of all the component parts of the car rather than having to stock them. This also means that each car can be customized to a much higher degree for each individual customer. These are the real advantages of bringing the customer into the loop.
Applications: The Final Layer
To date, the closest thing to a successful application for wireless devices has been e-mail. Devices such as the RIM BlackBerry have survived on this application alone, and some have called it the killer app for wireless. Actually it is more like the minimum ante for a wireless device. If it can't even do e-mail there is really no point to the device. To really get attention it must be capable of doing much more than that.
The question is exactly what that "much more" entails. There are those who are jumping directly to multimedia applications while others are building strictly vertical ones. The latter have the huge advantage that they can be built using much more flexible underlying networks, since they don't have heavy requirements around Quality of Service (QoS) or available bandwidth. That is why the first generation of applications will be much more about distributing corporate information than about the latest "Star Wars" release.
The key to building these applications is the availability of versions of programming tools like Java and Visual BASIC that are adapted to Web services. Note that these tools do not have to be specifically aware of wireless communications. That is the beauty of a layered architecture, and that is what will get these applications built while the wireless networks are still in flux.
Devices will start coming out this year that will be capable of extracting connectivity out of combinations of networks. They will utilize 802.11b when it is available and fall back to GPRS or CDMA when it is not. This will take place at the network level, creating the illusion of order for the applications. This is the only reasonable way to expect applications to be built, rather than expecting application programmers to be capable of programming around the nuances of wireless communications. Again, this is a major strength of 802.11b, since it usually succeeds in hiding the fact that it is a wireless protocol from the applications that use its services.
These devices will be important to applications programmers as well. Right now there is a definite demand for a device that cannot be completely described. We know that it is wireless, we know that it will either support a wide variety of applications or there will be a wide variety of devices that each support a small number of applications (maybe only one). We also know that laptop computers, cellphones, and PDAs all have strengths and weaknesses that somehow apply to this device. This year there will be a number of attempts made to define this device further, and most of them will fail. But with each failure we will come a little closer to figuring out just what will fill this particular niche until finally, the correct compromise is achieved. At this point that device will become an "instant" success.
People have been waiting for a thunderbolt to strike, the killer application that will suddenly have it all make sense. Those dreams were possible before computing was widespread, but today we have evolution instead. There will be a thousand small sparks, each of which will solve a small problem by applying wireless. One morning we will wake up and realize that the wireless revolution has happened around us. It is much less dramatic than paradigm shifts or new business models, but rather the way the world evolves.
Reader Feedback: Page 1 of 1
Latest Cloud Developer Stories
Subscribe to the World's Most Powerful Newsletters
Subscribe to Our Rss Feeds & Get Your SYS-CON News Live!
SYS-CON Featured Whitepapers
Most Read This Week