The real challenge with any new technology is sorting out the facts from the fiction and developing a firm understanding of 1) where the real value within the technology lies; 2) what the limitations of the technology are, and; 3) the promises the technology holds for the future as it gains adoption.
The perceived value of web services for desktop computers on the LAN has the industry speculating that thin-client computing will become a reality. The technology promises to simplify access to data and applications, thereby reducing enterprise application integrations and management/support issues that are time consuming and can cost corporations exorbitant sums of money. But what about the mobile environment? A steadily increasing number of users within corporations are using a wide array of mobile devices-laptops, PDAs, cellular telephones, tablet PC's etc. What value does Web Services offer to these users?
Special challenges for mobile devices
Part of the rallying for wide-adoption of Web Services is the switch from thick- to thin-client applications. For LAN users, some thought leaders think the time may be here for thin-client applications. But in the mobile environment, some barriers exist that have yet to be overcome.
Currently, there are three barriers to effectively switching to thin-client applications for mobile devices. The barriers are:
1. Loss of network availability: Mobile devices do not have coverage everywhere at all times. There are plenty of places - including remote locations, inside buildings and under bridges - where coverage drops off. Simply put; you cannot access special services if you cannot connect your device.
2. High connection costs: Wireless usage rates are getting lower but for a large enterprise they can still be very expensive. There are primarily three payment plans available: pay by the minute, pay by the amount of data, and pay a flat monthly fee. Any of these plans can easily add $60+ per month per user to a mobile deployment.
3. Latency: Wireless connections are slow. Even the fastest wireless connection is still slower than a dial-up modem connected to a LAN resulting in long delays between the time a user requests information and the time it takes to actually receive it. Services that take forever to operate frustrate the end-user and do not inspire adoption.
For years network providers have been promising high-speed, ubiquitous wireless networks that overcome each of the barriers listed above. But with national rollout plans for high-speed 3G networks shelved for lack of funding and with data and voice services competing for the attention of providers, the barriers do not seem to be coming down any time in the near future. Even if funding for advancements at wireless providers were limitless, the simple physics of providing continuous coverage make remote devices less reliable than those connected to a LAN.
Do these limitations mean that Web Services will not work for mobile devices or is it a matter of setting realistic expectations? Let's explore the instances where Web Services delivers on its promise for mobile computing and where it fails to deliver.
Promise #1 Integrating applications and data seamlessly across the enterprise:
Web Services promises to make un-integrated, non-communicating, discreet systems a thing of the past. In this instance, Web Services definitely delivers. For the enterprise system, server-to-server integration is a great fit for Web Services, which exposes computing capabilities through an API in a human-readable format. In this format, applications and data are portable across platforms and environments.
Today, Web Services is the easiest, most cost-effective way to get one application to talk to another application. Companies are writing web services wrappers around applications to make data more easily accessible. Additionally, product vendors are exposing web services for their software.
As applications such as Oracle, Siebel, and SAP start to interact more easily in this web services environment, new applications become possible. Imagine being able to have a single application that can access inventory records, price lists, and customer profitability statements without having to build huge data warehouses and without having to purchase yet another computer system.
The rubber hits the road, sometimes literally, in the mobile workforce. Mobile workers are out there in the field; they're making the decisions and collecting the data that keep their companies moving. The ability to access and input data into enterprise systems is crucial to these mobile application users. As enterprises begin to add software that communicates with mobile devices, they show a strong preference for vendors who already have Web Services exposed for their software. By making Web Services available in this way, vendors eliminate the need for their clients to have to write their own Web Services.
Exposed Web Services become the natural integration point that allows enterprise systems and mobile systems to take advantage of each other's services-no matter what development environment they're built upon.
Promise #2 Eliminating thick-client architecture for mobile devices:
Web Services promises more than just server-to-server integration. It also promises server-to-client integration. Because Web Services can be used to build thin-client applications, many companies are considering replacing thick-client architectures with lower cost thin-client architectures. This approach may be a good choice for LAN-based systems. However, while it is theoretically possible on mobile devices, such devices present unique limitations not found on the LAN. At least for now, the limitations of wireless connections-reliability, latency, and cost-prevent this promise from being fully realized.
To overcome the inherent limitations of mobile computing, companies must utilize two types of applications: 'connection dependent' applications for time sensitive data, and 'connection-independent' applications, for mission critical situations requiring that the data is available even if a connection is not. Most Web Services architectures assume that devices are always connected; as stated earlier, this simply is not true for most mobile devices.
Mobile applications need to be a blend of online and offline models, so the exclusively thin-client route isn't feasible. To illustrate the connection dependent and connection independent issue, consider this scenario: if a manufacturer sends a representative to a pharmacy with a PDA to take a product count and restock inventory, the rep cannot use his or her device inside the store because the store's security system and building structure interferes with the wireless signal. With a mix of thin- and thick-client architecture in place, prior to entering, the rep can connect in , download the necessary section of the database, enter the store, and perform inventory and replenishment work. Then, upon leaving the store, the rep can connect and transmit the data. If the device were strictly thin-client, the rep's PDA in the example above would be virtually useless once he entered the store, rendering him unable to perform his work.
Until wireless networks are as reliable, cost-efficient and fast as LAN networks, mobile applications still need to be a blend of online and offline models. However, this limitation represents the thick-client problems that IT professionals hoped to avoid-maintaining the thick-client and data on all the mobile devices is expensive and difficult.
Eliminating the cost for thick-client computing without moving to thin-clients
Nonetheless, with or without web services, it is possible to lower the cost of mobile application deployment and support to roughly the same cost as thin-client applications Using software to help actively manage these devices enables companies to lower the cost and increase the convenience of a thick-client architecture with capabilities such as:
· Software distribution
· Data backup
· Configuration setting
· Device security
Such features ensure that settings adhere to corporate policies, devices continue to work properly and data is backed-up to avoid losing user data. Furthermore, Device Management tools help companies protect sensitive corporate data on lost or stolen mobile devices.
Promise #3 Preparing and extending enterprise data for mobile devices:
In what is probably the biggest benefit for mobile computing, Web Services promises to help prepare and extend enterprise data for mobile devices, thereby decreasing the cost and time of integration. Web Services becomes the new middleware that allows synchronization servers and enterprise applications to work together without costly enterprise application integration work.
For CIOs, access to corporate data and application integration continue to be top concerns. Web Services meets these needs perfectly by creating an environment where applications on desktop computers and servers can talk to each other across the LAN. But the use of mobile devices adds another dimension to this integration challenge by adding new platforms and new end-users.
Web Services provides a perfect fit for integrating back office enterprise systems because they are always connected by relatively high speed, reliable, inexpensive connections. As these services are created for thin-client applications, they can be used or expanded for easily synchronized thick-client mobile applications. The basic premise is that Web Services becomes one of the interfaces through which a synchronization server can exchange data between mobile databases and enterprise databases.
An additional benefit from the lower integration cost is the ability to read and write to the business logic layer of an application as opposed to writing directly to the database. As a result, synchronizing data through a Web Services interface enables the application to validate the data writing it directly to the database. This ability to read and write to the business logic layer of applications helps companies maintains data integrity-a critical feature for any enterprise application.
By making the data required for a mobile application available via a web service, the traditional integration work done via proprietary API is minimized or eliminated, dramatically decreasing the cost of producing effective mobile applications. In fact, a great benefit of Web Services is the elimination of point-to-point integration. In other words, once we create a web service for, say a corporate portal, that same web service can be used for a mobile application that needs the same data. This benefit, combined with the easier method of integration, allows companies to leverage development costs across different projects.
The bottom line for wireless Web services
Overall, Web Services is likely to change everything about how enterprise system architecture functions-for both LAN and mobile devices. For corporations, the value of Web Services is clearly the cost and timesaving of avoiding extensive and costly integration work.
The limitations of Web Services for mobile devices rest mainly with wireless network connections. Many of these limitations may be overcome as wireless technology advances. For the time being, the limitations require a mobile solution that can handle synchronization and problems related to intermittent connections.
While Web Services may not solve every challenge surrounding mobility, nor allow enterprises to build adequately performing thin-client applications, it does solve two costly problems:
· Makes integration between applications on the LAN easier and more cost effective, which allows a mobile infrastructure solution to be added to the enterprise LAN infrastructure with little integration effort.
· Web Services prepares enterprise data for extension to mobile devices, allowing other technologies for mobile computing to work in a more straight forward method than traditional API.
Joe Owen joined XcelleNet in 1989 as the company's first product manager. Before his appointment as chief technology officer, Joe held several positions at XcelleNet including vice president of product management and vice president of business development. Previous experience includes serving as engineering manager at Software Link and software engineer on the IRMA development team at DCA. Joe earned a B.S. degree in Computer Science from the Georgia Institute of Technology's School of Information and Computer Science.