In-Depth

Editorial: Stuck in the Middle ... Again

• 06/01/2000

It’s an enigma, wrapped in a mystery, hidden in a puzzle. Call it what you want: product, strategy or philosophical phenomena, middleware has reared its many heads once again. But, is today’s middleware the Scylla of software, waiting to devour unsuspecting IT managers as they pass, or is it the end to a multi-faceted odyssey that finally, after a decade of searching, finds them safely at home?

Seven years ago – that’s a lifetime for many technologies – we reported on middleware coming of age. Even then, the term had been knocked around for nearly a decade, when it was first crudely used by some obscure software company long gone or long since acquired. Each year, middleware advocates promise that this will be the year of the "M." But, each year, the subject gets further clouded.

Seven years ago, the M-word covered translators and utilities, gateways and printer drivers, hardware and software. Today, as some users and vendors have finally come to grips with what middleware can and can’t do, others are still grappling to understand the concept and decide if they need it, while yet others have already moved on and abandoned the term altogether. Long-time middleware leader BEA, for example, no longer uses the M-word, referring now to "e-commerce transactions" as its latest buzz-phrase.

Today, the ability to transparently access data, in realtime, between mainframes and distributed systems remains the ultimate objective of IT management. Although achieving this objective is not likely to happen overnight, middleware hopes to speed the process. Four out of five experts agree that there are basically five types of middleware: Message Oriented Middleware, Object Oriented Middleware, Transaction Processing Middleware, Database Middleware and Remote Procedure Call Middleware.

Middleware was, and still is, like that little box of adapters one carries throughout Europe where power supplies and interfaces vary from country to country.

With the mainframe’s performance and security making it still the most popular method for storing corporate data, getting access to that data in a "user-friendly" manner remains crucial. And the challenge has only been amplified with the advent of the "open system" (remember that term) Internet front ends, such as NT, UNIX and Linux.

So our industry strives to bridge the two seemingly disparate worlds of mainframe power and distributed-systems flexibility in order to achieve "Universal Data Access." That’s why, this month, Pete Johnson describes an enabling technology that uses an OS/390 subsystem as the engine for high-performance access. According to Pete, such a solution "would operate transparently, taking native data; interpreting it and presenting it to distributed systems in industry-standard formats."

And we all know that getting from point A to point B isn’t always simple, especially when point A is on an S/390, and point B is somewhere on your enterprise LAN or the Internet. In our special report on middleware, Rob Innella outlines nine steps to better host integration to speed up projects, improve user satisfaction, and save money at the same time. And Mark Creamer describes a new class of software that has emerged, which promises to finally make good on the middleware moniker.

Has middleware defined itself? And if so, what exactly is its function? Before you delve into this issue, take another look at our front cover. I think it’ll start to answer the question.

The publishers of Enterprise Systems Journal, 101communications, will launch a monthly publication in August called Enterprise Linux. Enterprise Linux is looking for people to write on Linux security, clustering, application development, network management, database management, Linux vs. Linux distribution, and high availability and disaster recovery, as well as standards, training and education.