Are Mainframes a Dead End? Prove It!

Our readers comment on the October 2001 issue.

Dinosaurs Indeed?
I read [Linda Briggs’] recent editorial in Enterprise Systems entitled "Revenge of the Dinosaurs." [After spending] the past 20 years on large IBM systems, I’ve noticed that in most discussions on mainframe versus open system boxes, there isn’t much quantification of the data. For instance, in the Fortune 500 companies, what percentage use mainframes? How does this compare with five or 10 years ago? Quantification versus perception is extremely important for a serious discussion.

Also, a problem with mainframes such as IBM’s is that none of the universities seem to take them seriously anymore. I wonder if a significant number even have one or have access to one. If a computer science major comes out of college believing that mainframes are ancient technology and all they’ve been taught is Unix/NT on a small box, what’s the impact of this on the [mainframe] market? I would guess that it greatly affects it in a negative manner. No one wants to base one’s career on something perceived as a dead-end path. IBM could improve this situation by providing serious discounts for universities, along with funding for instruction.

John Parke
Senior Systems Engineer
Computer Network Technology Inc.
Marietta, Ga.

No Resurgence Here
In my humble opinion, there is no mainframe resurgence. What you may be seeing is a lifting of the fog and vapor. There may be an explosion of demand, but it will go unfilled. The young talent has been diverted and disillusioned. The skills needed come only with years of experience.

Look to Japan. Last I heard, they had elected to ignore the "client-server revolution" and stick to things that work over the long haul.

If any doubt remains, look at what is happening on the Internet. A few kids are now in functional control. To make matters worse, Microsoft plans to release even more powerful hacking tools with Win/XP ("raw socket" support). These tools will greatly reduce the skill levels needed to be a Internet killer.

The press must assume a large share of the blame for inventing terms like "mind share" and believing all of the hype, and then spinning the hype to make it sound like there was some substance [when there wasn’t]. Worst of all, it has ignored the facts.

Despite the fact that upper management has been in a frenzy spending money on networking infrastructure and hiring staff in those areas (with the usual lack of clarity as to what it’s for), we mainframers have continued to do more and more with less and less money and respect. Officially these legacy systems are static; in fact they’re forced to adapt, however tenuously, to every unplanned and unanalyzed change in the interfaces and in the management procedures (or non-procedure?). Frankly, it’s totally demoralizing.

Mickay Miller

Hybrid Security
I agree with [Roberta Bragg’s] comment regarding "when mainframes ruled supreme, it was security through obscurity…" If only management would appreciate the fact that most corporate systems are now a hybrid of virtually everything out there—thus compounding the whole idea of having a truly secured system. Looking forward to reading more from you.

Mark Ohiosikha
Bowie, Maryland

Thanks, Mark. I’ve got plans for some columns on security for hybrid networks. Glad you like the idea too.

—Roberta Bragg

In the Market
After reading Joseph McKendrick’s column in the August issue of Enterprise Systems, I began to wonder if he has any other documentation available. We are looking at several of the companies he lists and are trying to find out as much as we can!

One other quick question: Does Joe have any favorites among the group? We are looking at IBM, Jacada, ResQNet, Hummingbird, SEAGULL, Computer Associates, ICOM Informatics, Crossplex and Farabi. There are so many out there, I could probably look at them until I am blue in the face.

Chris Framel
Systems Analyst II

I've included a vendor list below. Your second question is tough, because all the solutions are fairly mature and consistent in their levels of performance. The main differences are the types of middleware architecture they employ—some can be run directly on the host system, others require NT servers. You may want to start off with a preview of IBM’s Websphere Host on Demand, and use that as the standard to measure the others against.

—Joe McKendrick


Anota Inc.:, Anota WEB-ifier

Attachmate Corp.:, e-Vantage Host Access Sever

Aviva Solutions:, Aviva for Java

Better Online Solutions Inc.:,, Jadvantage, Web-to-Host Emulator

BIS Advanced Software Systems Ltd.:, InterEST

CriticalPath (formerly PeerLogic):, InJoin PATH 3270

Ericom Software, Inc.:, PowerTerm Host Publisher

Esker S.A.:, Product: Persona Insight

Farabi Technology:, HostFront

GT Software:, Novation

Hummingbird Ltd.:, HostExplorer Web

IBM Corporation:, WebSphere Host On-Demand

ICOM Informatics:, Winsurf Mainframe Access

Jacada Ltd.:, Jacada for Java, Jacada for HTML

Looksoftware:, newlook

Microsoft Corp.:, Host Integration Server 2000

NetManage:, RUMBA 2000 Web-to-Host, RUMBA 2000 Web-to-Host (Host Java), RUMBA 2000 Web-to-Host (Host Express) Inc.:, Product: ResQNet, ResQNet Portal

Seagull Software:, J Walk, TigerRay

Standard Networks Inc.:, ActiveHEAT Host Access COM Object

Sybase:, XJS/390 Enterprise Integrator

WRQ Inc.:, Reflection for the Web, Reflection for the Web Professional Edition