Defining a Wireless Solution
5.3.3 Information Infrastructure
Information exchange is the underlying purpose of virtually every wireless solution. The data source, volume, and confidentiality requirements of a wireless application affect many aspects of the solution architecture, and implicate back-end application integration, data security, and network choices. The solution's information infrastructure supports the wireless application's information needs. For certain types of wireless applications, the information infrastructure is pre-supposed. Wireless Internet access or WLAN access takes advantage of pre-existing information infrastructures. Conversely, special-purpose wireless applications will likely need a separately defined architecture to handle information needs. For example, providing a field service worker with access to current customer status may require capturing and integrating information contained within various back-office customer and support databases.
The main drivers for information infrastructure requirements are the answers to the What and When categories of questions. In many ways, application functionality sets data requirements, especially from a source and content perspective. When a user needs access to data and how long that data remains useful define the mechanics of data transmission. Who is using the data and where it will be accessed affect considerations such as data formatting and display and security and back-up requirements. Like wireless applications, information infrastructure requirements are not easily captured through a simple checklist; however, the criteria in this section will help you develop a high-level understanding of your needs. As shown in Figure 5.8, the criteria in this section are: data sources, type of data access, volume, format, latency, immediacy, data integrity, synching requirements, and security. Refer to Chapter 12 for additional information on these topics.
The Information Infrastructure Cheat Sheet
- Data Sources This criteria applies to wireless applications that exchange data with corporate information systems and other repositories of information. A source may be a host database, file, a web site, or a wireless interface into a host application. Information from a single source may be delivered directly to the application, for example, a query sent to the wireless interface of a CRM package returns a response specifically formatted for the application, or it may have to be drawn from several sources and merged or processed before sending. In the latter case, the information infrastructure will have to contain the additional hardware and software needed to process the data. Processing may involve calculation, extraction, summarization, or reconciliation of disparate data to make it useful for the wireless application. As a starting point, list each potential data source needed for your application.
- Type of Data Access Data may flow in either direction between host and device. Data may be intended for viewing only, or for update (added or modified). A more complex architecture is needed to support data modification than simple data viewing. For each data source selected above, identify the appropriate data access requirements. For instance, an inventory tracking application may capture bar code information on the device for uploading into the corporate inventory database. A wireless banking application will allow customers to transfer funds between accounts. These types of applications require the ability to update host data.
- Volume The volume of information that the application will exchange has implications for data display, storage, and transmission. Given bandwidth constraints, data volume is an important consideration for network selection, but also affects application design. Application users may not tolerate the time and effort required to process and display large volumes of data, requiring creative design techniques to extract and present only the data most relevant for their needs. Volume of data ranges from very low, such as alerts and short messages, to very high for videoconferencing.
- Format Data format affects the level of processing needed to get the information into a form useful for the wireless device. Simple text data, web pages formatted for mobile display or files compatible with a given wireless application (e.g., a word processing document) may be exchanged "as is" without further processing. Other information exchange formats, such as XML and its variants, use stylesheets to allow the same data to be presented in different ways on different devices. This approach is used to "re-purpose" web-based data for simultaneous use on a wireless device. Given the limited data capabilities of many wireless devices, relevant data may need to be extracted from larger sources. At the high end, a complex query, such as providing a summary of sales activity across five territories, may require considerable processing to return the data to the wireless device in an acceptable format.
- Latency Data often has a useful "shelf life" or latency before it loses its value. Latency affects how and where data is stored and how frequently it must be refreshed to maintain its value. Some data, such as a list of inflation figures for 1990 through 2000, will never change and has infinite latency. Perishable information, such as the price of an individual stock during trading hours, changes on a moment's notice and loses value quickly. In contrast, an e-mail message may retain its value while it sits in an inbox for several hours. Estimate the latency of the major information exchanged by your wireless solution.
- Immediacy Immediacy addresses how quickly a given unit of data must be exchanged between the device and server. Stock price movements may require instant dissemination while inspection data may require only periodic updating. Serving the stockbroker with price information requires real-time processing and an "always on" network connection. The inspection application requires only hourly synching between device and server. Immediacy requirements affect both sides of the wireless exchange. A server-resident application that tracks packages may expect immediate notification of delivery problems from the field, but only periodic uploads regarding successful deliveries. Some server-resident information, such as stock price fluctuations, may require immediate dissemination while other information, such as last week's closing price, can wait until the broker specifically requests it.
- Data Integrity Data integrity assesses the risk to the organization if a piece of data is corrupted or dropped while in use by the wireless application. When data problems occur, if the symptoms are obvious rather than subtle, and if the data can be quickly and easily recreated, then data integrity is not really an issue. For example, if a download of the latest news from a web site is interrupted, the transmission can be restarted without concern for integrity. Conversely, executing a wireless fund transfer requires careful measures to ensure that the data is correct and reliable, and that all aspects of the transaction complete successfully.
- Synching Requirements Some data, such as a contacts file, may be maintained in multiple locations. Ideally, data should be identical at all times, across all locations, but this goal is not always feasible or practical. Synchronization is the method used to match data sources on a periodic basis. The period between synchronizations is determined by the latency of the data being shared and practical considerations for exchanging data. Synchronization may proceed in three ways. Data resident on a wireless device can be periodically delivered to a server. For example, a package delivery person sends collected delivery data to headquarters every few hours as he passes by a wireless access point in his vehicle. Server-resident data can be pushed out to a wireless device, replacing the version of the data kept on the device. For example, an engineering firm could send updated schematics to its field service workers' devices once a week. Or device-resident and server-resident data can be compared, matched, and merged to create updated versions at both ends, a process requiring a more complex synchronization scheme to deal with data conflicts or inconsistencies.
- Security The value and confidentiality of a piece of data determines the level of protection needed. Issues include who has access to that data, how it is transmitted and how and where it is stored. Authorization—access to a given function or piece of data—can be controlled at the information architecture level. For example, a user may be able to access personnel records from his department, but not from other departments. Highly confidential data may need additional security layers such as a secure network and/or encryption when it is transmitted. Given the high rate of theft and loss among wireless devices, a solution may require the encryption of confidential data stored on a device, or it may prohibit storing data on a device altogether.
Page 6 of 10