In-Depth

November Response Time

• 11/01/1999

I am a regular reader of ESJ. I find it very informative for my job. I have been working on the enabling of legacy systems to the Web. [In fact], we have developed our own tool kit to enable VT100 and 3270 screens to the Web.

Now, I am thinking of doing some experiments in enabling the client/server apps of Windows age. What I mean to say is: Why can’t we do a form scrape, like screen scrapping of 3270 (mainframes) and VT100?

Let me put it this way: I have a VB app in my Web server, and I want the forms of the app to appear in the browser for inputting the data, etc. The entire process should happen without touching my VB app in the server.

Now, I have some ideas of doing this, like taking help Terminal Server technology to push my screens to the browser for input. My query is: Has anybody done these kind of projects, and can they offer any advice?

ESJ welcomes all comments and ideas on the subject.

***

One comment on [Christopher Holley’s] August ESJ article ["SNA over IP – The Best of Both Worlds"] regarding DLSw scalability issues. These issues were resolved by DLSw V2 (RFC 2166) in 1997. Today, there are numerous examples of quite large DLSw network, including hundreds (and even thousands) of routers. If you are interested in additional information on DLSw V2 you can read the RFC, which, by the way, was authored by Dave Bryant from 3Com (formerly from IBM RTP). You can also find some additional information in the book I had the opportunity to co-author and co-edit with some folks from IBM two years ago: "SNA and TCP/IP Enterprise Networking."

You are correct that DLSw V2 (also known as DLSw+) does address many of the scaling problems which existed with the original DLSw implementation, and many large customers have been able to implement DLSw+ on a large scale by having a router farm of DLSw+ endpoints. However, there are some fundamental scaling issues which will always exist in a DLSw or DLSw+ environment:

DLSw Endpoints must maintain a full TCP/IP connection, as well as a full SNA connection and session.

In a Subarea Environment, the primary SNA boundary function must either reside in the 3745 or VTAM. If it resides in the 3745, then you run into scalability issues with the old NCP engine. If it resides in VTAM, then your scalability is limited by addressing limitations (only 64K Network addresses exist on a single VTAM).

In an APPN Environment, every DLSw endpoint is an SNA footprint, which means that all these nodes must participate in the APPN environment and absorb the overhead of broadcast searches, topology exchanges, etc.

***

I am totally unclear on the editor’s choice to publish the "18 Things to Expect for the Rest of 1999" in the "Countdown to Year 2000" column (September ESJ, page 73). I read each and every issue of this magazine simply because it is the best source of enterprise information available, but I am considering canceling my subscription based on the shotgun editorial techniques shown here.

Warren S. Reid is obviously setting a business foundation by confusing and terrorizing your readers, and I for one am not buying it. The half-hearted collection of un-truths which Mr. Reid is pushing was probably thrown together in a few minutes. It has no factual basis in the computer or business world, and serves only to worry or annoy.

Where are the facts from which these statements are being generated? Are we to rely solely on this soothsayer’s glances into the future? Two to three weeks would quite possibly bankrupt many small businesses and empty the bank accounts of many consumers. How do I know that? Like many of today’s working class, I live close to beyond my means, not because I like it that way, but because today is a time when both partners must work to make ends meet. It is a world where when a top executive makes a simple comment about the state of computer service business, the stock market responds with a noticeable drop. It is a time when we should do as much as we can to help those through what may be a long, hard road ahead.

Part of the function of a column is to include editorial opinion; and the opinions of Mr. Reid are based on countless hours of consulting, interviews and first-hand legal encounters. This was a prediction piece, intending to spark discussion, debate and, yes, even a little bit of controversy.

If you look back on past issues you’ll see that ESJ has fairly covered all aspects of today’s technology issues, most especially Y2K. Even Warren Reid has reported on and suggested in "Let’s Join Hands, Not Point Fingers" (July ’99) ways that readers can get through Y2K.

Mr. Reid’s September column is not the culmination or summation of ESJ’s Y2K reporting; but rather, it is a small portion of several years of reporting.