If your system is down, is your business closed? If so, the UniForum '95 track "Mission-Critical Open Systems" may be the right choice when you attend the show's main conference in Dallas March 14-16. The seven sessions of this track are designed for IS managers and support staff, and will explore the strengths and weaknesses of open systems for mission-critical applications, as well as provide guidelines for designing and integrating an open environment for high availability and recoverability.
The issue of reliable and recoverable databases will be probed in a session chaired by Mark Zucherman, technology program manager for the financial services industry with Hewlett-Packard Co. "Companies go to great lengths to make databases reliable," Zucherman says. "Once you have a failure in a mission-critical environment, the issue is how long does it take you to recover from that? You need to roll back the database and basically bring the system back up to a prefailure state. In the past that's been very difficult." For example, if the system is in the middle of a customer transaction during the failure, the database needs to be rolled back to a known state, and then the missing transactions need to be recovered. The panelists will try to address how long it takes to bring the application back on-line.
"There's all kinds of technology to make databases both reliable and recoverable," Zucherman notes. "The issue is to make recovery an easier process to administer, and a faster and more reliable process."
Zucherman's panelists will include Jeffrey Eppinger, director of product marketing for Transarc Corp., speaking from the transaction processing standpoint; John Pilat, vice president of the HP products division of Oracle Corp., speaking from the database perspective; and Jack Willis, co-founder and vice president of development for 1776, Inc., speaking about hardware integrity and reliability.
The issue of high availability vs. fault tolerance will be addressed in the session chaired by Wayne Fowler, director of technology network services in information systems for the Toronto Stock Exchange. "The intent of the session is to look at the choices available in the open systems marketplace for enhanced availability beyond the standard basic server box," Fowler says. Representing end users, Fowler will chair a session that features Bob Abraham, manager of system products for Stratus Computer, Inc.; David Ellenberger, division vice president of Aviion marketing for Data General Corp.; Jeff Tonkel, director of OLTP product management for Tandem Computers, Inc.; and Jim Johnson, chairman and founder of The Standish Group International, Inc., a market analyst.
There are basically three options in the market, Fowler says - high-availability systems that are not fault-tolerant, hardware fault-tolerant boxes, and software fault-tolerant boxes. Some companies offer one choice but not the others. "We are having some manufacturers of those types of boxes say why you would want to buy this stuff," Fowler says.
The session should be of great interest to users looking at developing or operating mission-critical systems on open systems, he notes. "You may want to put up a database application that's doing customer management, for example-anything like financial transactions, using open systems to run pieces of the business that have to be there," Fowler says. "I use UNIX systems to have the kind of availability that I need beyond the general high-availability systems. How much money do I have to spend to do that? The attendees at this will be people making technical decisions or building applications, and people running applications."
Other sessions in the track include:
For registration information to UniForum '95 you may: