Internet Systems' Frequently Asked Questions (FAQs)
What is Internet Systems, and for whom is it written?
Welcome to the premier issue of Internet Systems! As publisher, I'm proud to introduce you to this new publication and to the new world of corporate Internet computing.
Produced by Miller Freeman's Database and Client/Server Products Group, Internet Systems is the first Internet-related publication to focus exclusively on the needs of IS technical managers and senior staff who will soon be asked to deliver large-scale Internet-based, Internet-enabled, or Intranet applications. These IS professionals -- you -- are systems analysts, designers, developers, database specialists, DBAs, project leaders, and technical managers. Unlike Web surfing and programming tips magazines, Internet Systems covers the use of Internet and Intranet technology for strategic corporate applications: business-to-consumer, business-to-business, or intra-business.
Internet Systems will provide detailed technical information on application development strategies, tool selection, database access and data delivery, and integration with client/server and legacy applications. Internet Systems will also focus on opportunities presented by Internet technology for reshaping business applications.
Why not just call it "Intranet Systems?"
The term "Intranet" simply fails to adequately convey the scope of this technology. Intranet implies that an application runs behind a firewall, within the confines of a single corporate entity. We believe that as soon as you start to build Internet-capable applications, the nature of all your applications will begin to change dramatically.
In the design of corporate applications, every instance in which a customer or mobile user touches a system represents an opportunity to extend the software's reach. The building of this extended enterprise has the potential to change the nature of every application by allowing the software to "wrap" around the users and the customers. Even the most isolated "back-office" application will inspire reengineering to allow outside interaction. (Imagine a human resources application that lets you update your benefits package options from your home.)
What will I learn by reading Internet Systems?
Internet Systems will focus on tool selection, application development strategies, data access strategies, and emerging technologies, all within the realm of Internet computing. The publication will provide details on how to use Internet technology for data warehousing and decision support, customer service, electronic commerce, field force automation, information publishing, group productivity, OLTP, and other applications. Internet Systems will also track the progress of the software industry's key vendors in their quest to sell you on their technology. The race is on, and Internet Systems is your scorecard.
In deploying Internet technology, won't IS professionals be replaced by "Webmasters"?
No. Webmasters will have a continuing role in Web site administration; however, developing large-scale applications requires a team. That team includes specialists in systems analysis and design, application development techniques and technology, database design, database access, client/server and legacy integration, and middleware technologies. Internet Systems is written for the IS specialists who work in teams to design, develop, and deploy the corporate systems.
Also, the kinds of corporate applications that we cover in Internet Systems bear little resemblance to first generation HTTP Web sites that consist of static hyperlinked HTML pages and CGI applications. Internet Systems focuses on dynamic applications that employ Java, Visual Basic Script, or other development languages to provide rich user interfaces, rich data content, and complex processing.
Why should I consider using Internet technology for my corporate applications?
The three most compelling benefits of Internet technology are:
1. The ability to create customer-centric software, that is, to build applications that focus on customers as active participants. This can eliminate layers of reporting and can improve an organization's responsiveness. Remember when Business Process Reengineering (BPR) was all the rage? With Internet technology, BPR can now work to its full potential by enabling business modelers to wrap the software around the customers. Any process that involves a customer or a remote user can be optimized with Internet technology.
2. The ability to leverage the IT investment across a huge, distributed user base. An organization can deploy an application to one user, a group of users, an enterprise, or the entire world, without having to touch the client computers. The Internet browser becomes the universal client. With this kind of one-to-infinity scalability, an organization can increase by orders of magnitude their ROI in software development.
3. The ability to improve application content and to meet business challenges more quickly. These benefits accrue from the inherent scalability of the Internet architecture and from the recentralization of some computing resources. The scalability translates into a lower cost-per-user. Also, Internet computing is server-centric computing, which has great advantages in lower management costs. You've all heard complaints about the complexity and cost of client/server. Internet computing has the potential to reduce that pain.
Is Internet computing the death of client/server?
Client/server will exist for many years to come. First, it will continue to exist in dedicated internal applications. Second, in the Internet context it will integrate distributed resources for delivery to users via Internet servers. Nevertheless, the role of client/server may begin to shrink as a mechanism for delivering applications to clients. The scalability and the fact that IS does not have to configure and manage Internet clients makes a compelling case for developing new apps using the Internet architecture.
What are the key challenges of Internet computing?
With Internet computing, you must prepare for rapid and relentless change. When customers depend on direct access to corporate applications, you no longer have the luxury to schedule changes months later. If a line of business changes, the software must change immediately. If a policy changes, the software must change immediately. You can't mail out 10,000 sticky notes telling customers to "skip Item 5 on the main order screen" when the business no longer offers the "Item 5" service. Internet technology will force IS to align its development cycle with changes in the business. That's a challenge.
What about technical challenges? Can Internet technology do the job?
Today we have a rapidly evolving set application development tools. The biggest technical challenge is to create a development strategy that accounts for these changing toolsets. If your organization standardizes on Java, you must survey the field of third-party frameworks, compilers, and data access tools to identify which are likely to succeed in the market. If you invest in DBMS technology, you must consider how it will support future Internet applications.
The question of bandwidth will also dog the IS community. Today, remote users must access Internet services through LANs, WANs, public networks, and telephone lines. It is not clear when (if ever) the greater population of users will gain access to broadband communications that provide performance comparable to local area networks. The evolution of Internet applications will hinge on network performance. The sooner everyone has broadband access, the sooner the Internet will be able to deliver realtime audio, video, and other rich data.
Can the Internet handle online transaction processing (OLTP)?
Yes. But it will take some work. Transaction processing that involves multiple data sources or multiple tables in one or more locations still represents a challenge. Today's Web applications are essentially stateless. A user clicks a URL to view a page, and the server responds. The user then views the document locally, and may not communicate with the server again. This poses a challenge because many database applications depend on the successful completion of a series of forms in some linear fashion. An update to the database or databases must occur as a clearly defined unit of work. If any part of a transaction fails, then the entire transaction must fail. This kind of processing is problematic in a stateless environment because it requires an identifiable user who participates in a unique session. Unfortunately, a distinct user session implies that there is some overhead for each server connection.
Even assuming that you can avoid using the notoriously slow Common Gateway Interface (CGI) to access your back-end DBMS, Internet-based transaction processing is still problematic in terms of performance. High volume updating often requires TP monitors, stored procedures, or other forms of optimized and compiled database-updating code. Because the transaction processing code typically exists on an application server, Internet tools must accommodate application partitioning. (For example, a client applet loaded from a Web server could issue RPCs to yet another server that contains the transaction processing code.) Our product guide in this issue lists some Object Request Broker (ORB) technology to address this challenge.
How often will you publish Internet Systems?
Internet Systems is scheduled for May (this issue) and October 1996. In the October issue, we will announce a full publication schedule. We have not yet determined the frequency (quarterly, bimonthly, or monthly). If you like Internet Systems, please let us know!