(August 26, 2002) - As the worlds of wireless and Web services begin steadily to converge, Web Services Journal News Desk asks IBM's Andrew Donoho, Web Theorist, for a heads-up on where we are, where we're headed, and what technical challenges still remain in connecting the two technologies.

WSJ News Desk: For many firms, wireless devices have already made impressive changes in how business is conducted. Would you say we're deep into the transformation, or are we still just at the start of this revolution, with a lot further to go?
Donoho: We are at the beginning of the wireless transformation. We are just entering the transition to third generation cellular wireless - the first generation with any significant data capability - and are four years into unlicensed wireless (Wi-Fi - 802.11b) deployment. We are still working on deriving the true value of these data technologies and rapidly refining these applications. I would also say that wireless is affected by the maturation of other technical trends and by the limits of the human interface to small devices. For example, Web services technology is being driven by its ability to reduce the costs to integrate disparate systems. Wireless application deployment will be subject to these same economics. From a human interface perspective, there are limits to what can be displayed in a small, battery-powered device. In practice, the limited size of displays will affect how useful these devices can be.

WSJ News Desk: IBM has been developing its strategy regarding Web Services in relationship to cell phones, PDAs, pagers, and other devices that need to communicate with each other. What are the biggest hurdles to making Web services work on devices?
Donoho: There are several challenges to using Web services on devices. The first challenge is finding the device. For example, cell phones are designed to be found by the network. However, how can a home wireless voice over IP phone be found on the home network, or on the Internet? There are multiple, loosely consistent protocols vying for market acceptance in the area of local network discovery - SSDP (Simple Service Discovery Protocol) and ZeroConf (Zero Configuration Protocol) are two examples. There are many more in various niche areas that are using TCP/IP technology.

The second challenge is definition - gaining consensus on simple, common services. Although Web services are dynamically described, they will still need to be defined. For example, if you want to control your home MP3 server to load a song into your combo phone/PDA/MP3 player, then some control protocol needs to be defined. This is why organizations, such as the Web Services Interoperability (WS-I) Organization, have been created - to ensure interoperability between implementations of defined services. Those common services need to be defined. Two organizations are currently addressing this area - Open Mobile Alliance (OMA) and Universal Plug and Play (UpnP). We expect more organizations to start defining these services for their niche applications.

The third challenge is description - how do you advertise a device's services? For example, UpnP has an XML description document. An alternative would be the Web Services Inspection Language (WS-Inspection) that IBM co-developed. WS-Inspection defines how an application can discover a Web services description on a Web server, and it allows any device to advertise its Web services in a generic way.

Media devices on Firewire (IEEE 1394) have yet another mechanism. These mechanisms need to get sorted out. In classic Internet style, we will probably end up with several active protocols for description.

WSJ News Desk: How is IBM approaching the security issues of doing business on wireless devices?
Donoho: Security is a foundation technology for doing business on any network. IBM is helping to define the WS-Security protocol in OASIS with many other companies. We are also actively engaged in understanding the security issues that affect the wireless community and communicating that back to OASIS. Some of these security issues are solved in different ways by different device applications. For example, European cellular phones use a tamper proof SIM chip to prove the identity of the phone and, hence, user. Identity may be determined using a different technology in a home security system (PIN) or a Wi-Fi enabled laptop (WEP Key). It will take some time to cover the wide range of device applications, probably more time than in the relatively homogenous world of server transactions. Because of this diversity, WS-Security has been designed with pluggable security technology in mind.

WSJ News Desk: What is IBM doing to help revise existing protocols (such as SOAP or WSDL) or to develop new protocols to enable messaging and communications on these smaller devices?
Donoho: The key issues with wireless devices and network protocols are size of the message, the error rate of the wireless channel and latency of the connection. The size of the message is controlled by its format as it is serialized and placed on the wire or via radio into the air. Binary messages can be more compact than XML messages. Because of this, SOAP 1.2 supports multiple kinds of serialization of messages - both XML and binary DIME serializations exist today. Other formats may exist tomorrow. There is active work on this issue in many venues. IBM's leadership in the W3C helped enable the flexibility of SOAP. We will fully support both XML and DIME encodings. We will evaluate other encodings as customers and business conditions warrant.

WSJ News Desk: What improvements is IBM working on to help Web services on devices better serve the salesforce and workforce ?
Donoho: One of the key issues with salesforce automation is, "how do I push the right data to the salesman's device?" If we have to depend upon device pull, then wireless devices, not just wireless web services on devices, will not meet the needs of the sales force. Salesmen are not data administrators. The data they need must be available in a timely fashion. Therefore, we need to remotely store data on the device. SyncML is one such technology for doing that. The Web services community may develop other services to remotely manage these devices. It is still early to tell in what direction the industry will go. IBM is continually working with software providers and device vendors to help shape these protocols.

WSJ News Desk: Will Web services on devices ultimately have a greater impact upon large corporations or upon small and middle-range businesses?
Donoho: Web services is about gaining efficiencies by integrating systems. To the extent that a business uses automation systems, then Web services should reduce their cost to deploy and maintain these systems. Therefore, I think the use of Web services on devices will primarily help medium and large sized businesses. The one exception would be if the Application Service Provider (ASP) market matures. Then small businesses could also benefit by subscribing to an ASP web service for their workforce.

WSJ News Desk: With the difficult economic climate lately, businesses may be reluctant to invest in new strategies. What advice do you have for companies who are skittish over new technologies right now?
Donoho: Web services has a different technology proposition than most new technologies. It is focused upon reducing the cost of exploiting your existing automation technology in new ways. Because it is based upon XML, it offers the real chance of allowing integration to extend outside of the enterprise. This should result in new business efficiencies and opportunities. Web services gives you better life cycle costs for deploying technologies.