In-Depth
Managing the Output...IT's Last Headache!
As companies come to rely increasingly on distributed applications foressential business processing, IT managers have gone to great lengths to keep theseapplications up and running. However, most have yet to address the equally important taskof managing the output generated after processing has been completed. Output problems nowaccount for 30-40 percent of help desk calls in distributed environments, on average, andmanagers are increasingly coming to realize the importance of effective output management.
In this article, we'll explore the issues of output management in the distributedenvironment, identify the challenges and suggest solutions.
Output in the Distributed Environment
A typical large organization generates hundreds of thousands (frequently millions) ofpages of output per day, much of it business-critical, and all of it important to the userwho requested it. Output flows from a wide variety of applications ranging fromdesktop-based spreadsheet and word processing programs to specialized workstation-basedgraphics applications to the massive, UNIX and NT-based Enterprise Resource Planning (ERP)applications (from SAP, Baan, Peoplesoft, Oracle, and others) that run many businesses. Inaddition, most large organizations still generate a significant amount of output fromtheir host computers.
For management purposes, output can be divided into three categories:
- Production. This includes the reports and documents generated by production applications: reports, invoices, packing slips, statements, paychecks, work orders, etc. The key management challenge here is to ensure efficient, secure, and reliable delivery of critical business data to the output target, and to automate repetitive output jobs.
- Commercial/Reprographics. This includes all internal publishing of manuals, documentation, marketing materials, etc. Here, the key management challenge is ensuring the integrity and efficiency of high-volume print jobs. (If a printer goes down halfway through a 50,000-page job, restarting is not an appealing option...)
- Distributed/End User. This includes all output generated or requested by end users, from word processing documents and spreadsheets to e-mail messages and documents requested from the Web or online archives. The key management challenges here are providing transparent access for end users to print to other output devices anywhere on the network for which they are authorized, and notifying them of job status.
These categories provide a useful way to think about output. However, in practice, thelines between categories frequently blur: A company running an ERP application, such asSAP R/3, may generate a steady stream of reports, forms and other documents on a dailybasis (production output); this output may be consolidated in the form of a quarterlyreport which must be printed on high-speed printers and sent in hard copy to the entirecorporate sales force (a reprographics task).
The report may also be posted online, and individual end users may request parts of thereport to be printed on a local printer, or faxed to a remote user in another country(distributed output).
Thus, while each category of output requires a management solution, a corporate outputstrategy must address all applicable categories to provide the maximum benefits.
Output Management: The Challenge
In the distributed environment, output can be generated from applications runninganywhere in the corporate internetwork, on any of a range of platforms, and in volumesranging from a single-page memo to a 50,000-page high-speed print job. The output may bedirected to any of a range of output devices (printers, fax machines, digital copiers,pagers, e-mail addresses, Web pages, archives, online report viewers, etc.), locatedanywhere in the organization, or outside the organization with customers or suppliers.
IT administrators face the challenge of ensuring that all this output is deliveredreliably, efficiently and securely to its final destination.
Unfortunately, most distributed applications and OSs offer very little, if any, outputmanagement functionality. ERPs, for example, typically classify print jobs as"successful" when, in fact, they have merely been spooled. OS utilities don'tprovide the visibility or auditability of print jobs required in production environments.Utilities provided by output hardware vendors offer only narrowly-focused management ofspecific devices.
Output Management: The Solutions
Currently-available automation tools can greatly facilitate the process of managingdistributed output. However, many large IT organizations, desperate for a solution,purchase tools without taking the time to understand the way output is generated and usedin their specific environment. Just as an ERP project must model corporate data flow, anoutput management solution, to be effective, must model and support the output flow of anorganization.
Depending on the environment, managers may look for tools, such as Tivoli's Destiny orIBM Printing Systems Company's InfoPrint Manager. These kinds of tools offer functionalityin the following areas:
Optimization of Printer Usage. With multiple laser printers on every LAN,high-speed printers in the central office, and a range of other output devices scatteredacross the network, most large corporations have far more output capacity than theyrequire. However, they have no way of marshalling output resources intelligently tomaximize overall speed and efficiency; as a result, they continue to purchase unnecessaryprinters and other output hardware. Management software can improve resource use byintelligently routing jobs to resources based on their characteristics (size, destination,confidentiality, priority, etc.) and the status of print resources (balancing job loadacross available printers, for example). Software can also help by maintaining up-to-datestatus information on output devices and by scheduling maintenance.
Visibility/Auditability of Print Jobs.
In the unmanaged output environment, users and administrators have no way of checking thestatus of their print jobs. Centralized management tools enable administrators to view thestatus of all output jobs and the location and status of queued jobs. Managers canlogically manage print jobs, moving them between queues, for example, if a printer goesoff-line. They can also request that notification be sent to users or administrators whenjobs have completed. In some cases, end users can access the tool to check the status oftheir jobs, without bothering IT adminstrators.
Security. For production output, adminstrators must be able to enforce corporatesecurity policies, limiting user access to sensitive output and enabling users todesignate print jobs as secure, or direct them to print on paper marked"confidential." Output sent across the corporate intranet or the Internetnetwork may also need to be secured against unauthorized access through compression andencryption.
Fault Tolerance. Critical output jobs (such as those generated by ERPs) must beprotected against failure of an output device (fax machine out of paper, printer down formaintenance), a server or a network segment. Functionality is required for automaticallyre-trying jobs, recovering failed jobs and re-routing jobs sent to failed output devices.
Automation. Most large organizations have complex, repetitive output jobs (forexample, sending a sales report to all regional managers, posting it to a Web site,notifying the Sr. VP of Marketing by pager of its availability, and archiving a copy). Inthe absence of automation tools, these tasks are likely to be carried out by thelabor-intensive, error-prone and time-consuming distribution method known as"sneakernet." Automation tools enable administrators to pre-define complex printjobs, so that distribution occurs without administrator intervention. Such tools may beintegrated with the application generating the output (typically an ERP) for greatercontrol.
End User Empowerment. No longer content to wait for their host-based green-barreports to be printed and shipped from the data center across the country, end usersincreasingly demand the ability to request output in specific formats (i.e., via e-mail orsent to a remote printer). In addition, end users may need access to print resourcesaround the network from their PCs, and the ability to automate repetitive output tasks(establishing e-mail group lists, for example, or print distribution lists).
Format Conversions. Applications generate and output devices require output in a range of formats: IPDF, PCL, Postscript, AFP, CCIT (used by fax machines) toPDF, and many others. Frequently, a document must be converted from one format to anotherfor output to a specific device (for example, if an R/3 report must be sent to a faxmachine and an e-mail address, the same PCL document must be converted to CCIT and PDF,respectively).
Mainframe Connectivity. Where the mainframe is still used for corporate dataprocessing, output must frequently be sent from the host to end users via the network.Adminstrators may also want to leverage high-speed print resources connected to the hostfor jobs generated by distributed applications, such as R/3.
Management of Specialized Print Resources. High-speed printers, digitalprinters, color printers and other specialized output equipment all require a high degreeof administration to operate correctly. Typically, these output devices come with theirown specialized management software. However, administrators may also wish to tie theminto the overall output network, so that they can receive jobs from network applicationsand output jobs for delivery across the network.
Online/Web-Based Output. For many organizations, publication of output online(in a user-accessible archive or to a Web page) can reduce the cost and waste of hard-copyoutput, while giving users convenient access to reports and other corporate data. Someonline report viewing software lets users perform searches against the data, creatingreports from the posted information.
Imaging. In many cases, organizations wish to store images of output, ratherthan merely the data it contains (credit card statements, for example). Imaging systemsenable users to quickly search for and recall specific documents from an online archive.
Output management should be addressed as early as possible in the transition todistributed processing. Particularly in ERP roll-outs, IT frequently waits until afterroll-out to confront the problem, only to discover that the effectiveness of theapplication is being seriously compromised by the difficulty of moving output.
About the Authors:
Andy Kicklighter is the Product Manager for Output Management products at TivoliSystems. Randy Larsen is the Worldwide Product Manager for IBM's InfoPrint Managersoftware. He can be reached at [email protected].
ManagingA BI Primer
Gaining Insight into the Business of Doing Business
By Rob Rose
The focus of Business Intelligence (BI) is strategic, rather than tactical. Businesspeople use the results of their BI efforts to make the kind of decisions that impact thebottom line, decisions that lead to cost reductions or revenue generation.
For years, organizations have been attempting to empower workers with decision-makingauthority in efforts to make the organization more responsive to the marketplace and thecustomer. This, however, requires that empowered workers have information readily andwidely accessible a phenomenon dubbed the democratization of information. Butinformation alone doesn't lead to effective business decision making; effective decisionmaking requires BI.
Through good BI, IS insulate users from technical intricacies and physical structuresof the data. Instead, BI presents information in the context of the business, its rulesand its structure. As a result, regular business people don't have to wrestle with thetechnical issues surrounding data, databases, structured queries and such.
As part of the mainstreaming of information, the BI solution handles the management andcontrol of the information. It also handles the security aspects. All of this occursbehind the scenes, freeing users to explore the information for business insight withoutworrying about data management and security issues. Finally, BI involves a coordinationand sharing aspect. Business people need to share insights gained from BI and coordinateactions taken on the basis of those insights.
The BI Effect
BI is often confused with one or another technology: OLAP, data warehousing, datamining, ad hoc query, and others. These technologies and software tools based on themfacilitate the delivery and presentation of BI to users.
BI-enabling technologies collect and consolidate data in data warehouses and datamarts, store and manage data in relational and OLAP formats, and generate SQL queries.However, the role of these enabling technologies is to make information available so thatthe BI effect can happen. They are not a BI end unto themselves.
Similarly, BI differs from other technology solutions, such as decision support systemsand reporting systems. These systems typically provide operational information that helpmanagers make decisions and help workers in the ongoing performance of their varioustasks. Or, they provide status reporting where a situation is at a given moment. Thisleads to the BI effect.
The BI effect is really about business performance. The answers users seek from BIsolutions address how the business is doing, where it is lagging and where it hasopportunities. This insight can often be captured in a series of key performanceindicators. Managers can define key performance indicators for every function andoperation of interest and use BI solutions to track and analyze these indicators overtime.
But the BI effect isn't limited to one type of activity or group of users. The valuethat BI delivers is very broad. It begins when people simply view reports that revealsomething about their business. From this simple exposure to information, the value of BIextends all the way to robust analysis by which people use patterns in information toforecast business results.
The BI effect is further enhanced through the ability to share insights andconclusions. BI results can be reproduced and disseminated to appropriate individualsthroughout the organization. This sharing leads to collaborative action and directed,coordinated execution that improves business performance.
Critical BI Issues
The following issues must be addressed in any BI solution:
Standards. Support for both client/server and Web computing, support forstandards that facilitate interoperability with the existing computing and datainfrastructure, such as ODBC, standards that facilitate interprocess communications andsecurity. Other standards will address the sharing of metadata, business rules and datastructures. This also entails a commitment to open standard interfaces, as well as theadoption of emerging standards, such as Web directory management, XML/DHTML andvendor-initiated standards OLE DB and SAP's BAPI.
Integrating BI applications directly into the data warehouse. Approaches thatleverage the availability of data warehouses, data marts, ERP systems, metadata sourcesand security models. The objective is to create an enterprise environment that allows theseamless sharing of common administration, security, business rules and metadata among BItools and applications.
Emergence of a low cost BI infrastructure. Support for lower cost platforms,such as Windows NT, MS SQL Server 7 and OLAP Server. While Microsoft triggered thiseffort, other vendors have followed with cost lowering initiatives. This produces tworesults: It allows mid-size and smaller companies to participate fully in BI, and it leadsto a proliferation of departmental BI efforts, which are now within reach of departmentalbudgets.
Emergence of BI portals. Following the popularity of Web portals, BI portalsoffer the aggregation and categorization of information found at Web portals in thecorporate BI environment. This allows for the creation of enterprise information portalsaccessed through a robust, Web-like interface.
User role coverage. Ability to meet the needs of all users, from those requiringreporting and simple query capabilities to those who want to explore information in depththrough data mining and sophisticated OLAP analysis.
Completeness. An end-to-end integrated enterprise approach that begins with theplanning for the initial data mart and continues all the way through the receipt ofmeaningful information by users.
The successful organizations in today's global information-driven economy will be thosethat use BI to track and improve the performance of the business. In such organizations,managers will define and monitor key performance indicators. They will drop down toexamine the detailed data behind situations where indicators suggest problems oropportunities. And, they will share the insights they gain with other workers and managersas they coordinate strategies and execute initiatives to improve the business.
About the Author:
Rob Rose is Vice President of Product Marketing at Cognos Corp., and is responsiblefor strategy and positioning of all Cognos products.