An Overview of Wireless Computing For Enterprise Infrastructure
Those of us who have responsibility for creating the systems that power corporate infrastructure have been inundated with articles regarding wireless technology. Does this technology have a place in business today? If it does, what is this role and how can we use this new technology sensibly? Much of the current crop of solutions seems like technology searching for an application. There are a number of examples of consumer wireless applications, but fewer examples exist in the context of enterprise infrastructure.
Almost all enterprises have mobile workers. These include, sales forces, field service representatives, delivery people, and other people who go where the customers are. Other mobile workers, such as team leaders and project managers are less obvious as mobile workers, but still need to be supported away from their offices. Quite frequently executive staff members qualify as mobile workers as well. Reconnecting these people to the internal systems of the corporation also improves customer service by improving the quality of their information.
Wireless technology does have a role to play in our enterprise computing strategy. Wireless applications improve productivity by improving corporate communications. Properly designed wireless systems can be used to shorten billing cycles, enhance our dispatch and logistics systems, reduce the cost of outfitting mobile workforces, and that’s just the tip of the iceberg.
Current technology can easily support the goals I’ve mentioned. The capability exists for any enterprise to implement systems that serves these goals. In fact, we can go quite a bit further than this if we wish to, using off the shelf technologies like XML, WML, two-way paging, and wireless modems. Mobile Point Of Sale (POS) terminals are easy to integrate, and data gathering devices can be purchased for next day delivery.
Although some uses of the technology are easy to envision, like integrating contact lists to cell phones to give the sales force current customer data, others are less obvious. A less obvious example might be automating the expense reimbursement process. This sort of system requires numeric entry, and selection of items and categories from lists that don’t change much. If someone has an unlisted expense then there’s a small amount of text entry required. Mobile systems positively identify the user so the only additional requirements are management control options and a standardized method of gathering supporting documents and receipts. Applications with these features work well when using wireless technologies.
The easy way to discover good candidates for mobile applications is to look at all of the points of contact between a mobile worker and enterprise systems. This will show almost all of the obvious applications. Finding the less obvious uses is a more difficult task, but this is how the less obvious but larger productivity gains are often to be found. Interviewing mobile users to assess their needs is another good way to find infrastructure gaps.
Any communication between a mobile worker and an enterprise system can be a candidate for wireless systems. But every device can’t support all applications. Display restrictions, text entry difficulty, and other factors may require that some applications be implemented only on devices that are appropriate. Wireless systems shouldn’t be implemented as a simple translation of current web technology. Web technology is navigation intensive and the delays encountered while browsing will be severe. Wireless systems require careful implementation and very focused goals. Much of the disappointment in current wireless systems stems from implementations that have created slow cumbersome versions of existing web systems. A mobile worker needs to be able to get in, do whatever the task at hand requires, and then get out, and get back to being productive. Browsing is not a mobile activity.
The requirement for training is often overlooked with mobile systems. Part of the problem is that it is difficult to take many of the mobile users out of the field for training. There are attitudes that must be changed to ensure that the mobile devices are not just used in the traditional manner. Most people know what cellular phones and pagers do, and when we make them do more than that we must show them that the device has changed. Phones that do two-way data transmissions, or pagers that can run a program are beyond the expectations of new users. When designing the system, it may be advantageous to bring members of the user community into the design process.
Many of the packages that form our wireless infrastructure are beginning to have wireless modules that can be added on. Unfortunately, this approach often doesn’t really suit the needs of mobile workers, because these add-on pieces are not integrated into the workflow and business logic of the enterprise. True “turn-key” solutions do not currently exist, so organizations without experienced wireless developers may find it difficult to recruit or train staff to create and support these new applications. Although the tools for creating wireless infrastructure are improving, a true rapid application development environment for use by intermediate to junior level programmers does not currently exist.
The same questions that must be answered to build a good application must still be answered in the wireless space, but there are some additional concerns that must be factored in as well. The scope of the deployment will dictate what technologies can be considered. A national system will need to use different technologies than one that operates in metropolitan areas, as the digital infrastructure for cellular communication is not built out in more rural areas. Many of the digital communication methods have a delay in presenting information that impacts the design of applications. Technology change must be managed as well because of the rapid evolution in this area.
When examining mobile workforces there are specific needs that must be addressed. Some of these needs vary based on the role of the mobile worker, and on the vertical market. Generally speaking these needs include voice communication, notifications of various kinds, data entry, and data retrieval. These functions are typically parts of existing systems in the enterprise such as dispatch, logistics, sales, and support. Assessing the needs of the mobile workforce is critical to integrating new infrastructure to support mobile workers because poorly designed systems can produce new inefficiencies instead of benefits.
There are a lot of mobile workers with minimal resources available. One example of this is a worker that is moving about on foot or via public transportation in an urban environment. This worker has only the things that they can carry with them. When designing mobile support infrastructure for this worker we will be limited to devices that can operate for extended periods without being recharged, or that have lightweight power cells. The number of devices this worker can be burdened with must be limited. Systems supporting this worker must be tightly focused on the work to be done, immune to things like heat and cold, and useable in both brightly lit and low light conditions.
In addition, we need to understand the contact frequency that this type of worker will require. A good example here is the contrast between a data gatherer who may collect a large number of records without needing to make contact with the enterprise as contrasted with a person carrying a portable point of sale terminal who will potentially need to contact the enterprise every few minutes to validate credit cards and commit transactions. Since most of these portable devices spend a significant portion of their power budget on transmissions, these two applications imply significant differences in the hardware supporting the worker.
Another type of mobile worker works from vehicles. These workers have access to a power source from the vehicle’s accessory circuit and can also carry documentation and some peripheral equipment. Communication infrastructure may be also be built into the vehicle. This sort of worker may need to leave the vehicle for extended periods, so the previous model may apply to them for some percentage of their working day. In these circumstances it becomes necessary to provide support for use of more than one device to communicate with the mobile worker. This introduces routing requirements that may be non-trivial. In some cases, this sort of worker may also require storage of messages that they may not be able to respond to when working out of the vehicle.
Many workers with access to wired resources also spend time out of their offices. These workers spend some percentage of their time in a mobile mode, and they become wired users when they are in their offices. In this model integration between the existing enterprise and the new mobile infrastructure is a crucial feature, as the workers will not want to have to spend valuable time synchronizing their devices.
When designing infrastructure to support all of these workers it is important to understand and support an efficient workflow. New systems do not come into existence in a vacuum. We can’t succeed in adopting wireless technology without recognizing and leveraging the existing infrastructure. A properly designed infrastructure will extend the existing systems without requiring massive changes to the operating infrastructure.
In addition, mobile applications are different than the normal fixed site systems in that they run on hardware that has limitations. These limitations vary from device to device. The limitations include screen resolution and size, restricted user interfaces, memory constraints, and power limits. There are also message size limits and network latency with some technologies.
Networked non-mobile systems tend to run on multi-tasking workstations so that application switching happens without any serious latency. Many systems manage these context switches so naturally that the end user may be unaware that they are not using a single application. Providing applications that can communicate with multiple back end systems is critical to user acceptance of mobile applications as the mobile infrastructure has latency that can make switching between applications tedious.
Screen navigation must be carefully thought out, and quite frequently multiple back end system’s data must be presented together in a single wireless application. Great care must be taken to insure that the mobile application supports the workflow that the mobile worker needs. The user interface available on a given device will strongly impact what application features should be presented. Some devices are better for text entry, and others will lend themselves more to entering numbers or picking options from lists.
The types of mobile devices fit into specific categories. These include Personal Digital Assistant (PDA) devices with some sort of data connection, data enabled telephones using Short Message Services (SMS) or Wireless Application Protocol (WAP) technologies, two-way and one-way paging equipment, special devices (like those point of sale devices found in the car rental business), traditional wired lines (which are not mobile, but the user may be at a kiosk and be mobile). This gives us an environment where users may be entering our system via a highly variable set of entry vectors, each of which has variation in display capability, connection quality (robustness), control sets, and infrastructure support. This kind of an environment can be challenging to administer.
We have taken a look at some of the issues surrounding the development of wireless infrastructure for enterprise computing. This is a rapidly evolving field that has great promise in the near future. Carefully designed and implemented wireless systems will enhance our operations by providing more efficient communication and by making existing systems more streamlined for our mobile workforces.
Please tell me what you think about this content and how I might improve it.