Friday, May 20, 2005 (A.G): A Plug For IBM's CICS Transaction Gateway

This is what I was hoping to cover yesterday, as an adjunct to the "Battle Royal for Legacy Capital" Part II that I had posted earlier in the day.  But then I got, justifiably, side-tracked with that galling 70% claim.  Well I saw it again this morning in a PPT presentation.  But the presentation was from Canada and the claim was attributed to Gartner (albeit not Gideon per se).  Well, I guess that just about puts the lid on that matter.  SMILE.  A few years ago I would have, automatically, made a joke about what 70% of anything is worth in Canadian dollars but the US dollar, alas, is not what it used to be.

Anyway, the good news is that when it comes to CICS, there is no doubt or debate as to the fact that it is, incontrovertibly, the #1 transaction processing system (or monitor as we used to call it before that term got waylaid for other uses) in the known Universe.  The granddaddy of them all.

CICS preeminent role vis-à-vis transaction processing dictates that any enterprise class EAI solution, if it wishes to be credible, has to have a story as to how it interfaces with CICS.  Consequently there are a lot of CICS adapters out there, running on a variety of platforms and using a bewildering array of APIs and protocols to interact with CICS.  If CICS transactions are going to play a major role in your new SOA-based composite applications you need to make sure that you are using an efficient and scalable CICS adapter.  Don't just assume that all CICS adapters are much the same and just variations on the same theme.  Not so.  Some will get you CICS transaction but will do so extremely laboriously.  There is at least one, that I have had the misfortune to see, that uses a dual gateway; i.e. back-to-back adapters.

So be careful.  Make sure you thoroughly check out these mainframe adapters.  Look at IBM's CICS Transaction Gateway as a benchmark -- both for architecture, approach and performance.  There is also the CrossPlex solution that has been discussed here.  That is it for this week.  Enjoy the weekend.

Thursday, May 19, 2005 (A.G): We Have To Stop Using This Hokey 70% Number

Given that I just posted Part II of the "Battle Royal for Legacy Capital" I was going to use today's BLOG entry to plug IBM's CICS Transaction Gateway because all of you looking at CICS adapters should at least take a look at this option.  But doing some background research (and yes, I do do some intermittently) I found an article, written in 2004, that started with the infamous: "... with more than 70% of the world's data stored on mainframe computers ...".

I couldn't let that pass.  It is too embarrassing.  We have to STOP using this 70% right now.  It is so wrong that I find it hard to believe that anybody can still say it without realizing that they are being naive, if not outright stupid.  If you add enough caveats to it, you could just about justify it -- but to claim 70% of the world's data.  That is just plain wrong.

Before we even get to PCs and Unix servers, lets just stop and think for a second about all the AS/400s out there.  Now if you weasel word it along the lines of "Fortune 500 corporate data" you could be getting close.  But just to say 70% is crazy.

InfoWorld has already identified that this 70% is an IT myth.  I had admitted a long time ago that Guy Hoffman (the guy who made OCS famous) and I, fed this number to the media and analysts, 1996 -- 1997, at a time when people took note of what OCS said because they were said to be #1 in Web-to-host.  Even in 1996, we took the precaution, often, of adding the "large corporation" caveat.  But even that was a stretch.  It has always surprised me that the Unix camp has never refuted this claim.

Wednesday, May 18, 2005 (A.G): Gridlock At Sun

Some of you may find it hard to believe but I really like Sun and admire Scott.  Honest.

A long time ago, around 1992 - 1993, I even used to bum around with Scott's younger brother (of 18 months) who was the V.P. of sales/marketing APT -- where APT among other things had one of the original LAN-over-SNA products.  I was repeatedly told that Scott is an outstanding guy, who gave large shares of Sun stock to all his family, and though very rich and famous, even in the early 1990s, liked nothing better than to kick back with a beer and watch hockey.

I just worry about Scott because I really don't want him to go too far astray or even worse become Bill's lap-dog.

So it really grieves me when I hear that Sun has been forced to delay the availability of its imaginative, $1/CPU hour, utility computing 'Sun Grid' scheme.  Now if I didn't like Sun/Scott I would now go onto mock them for their rather naive excuse.  But I won't do that because I like them.

So what is their excuse?  Lack of computing power!

OK.  No tittering.  Be nice.  They have been under a lot of stress.

The irony!  Grid computing initiative stalled due to lack of computing power.  I, noted for my infectious giggle, could start giggling.  But, poor Scott.  Aahhh!  Lack of computing power?  Isn't grid computing ALL about harvesting currently UNUSED computing power.  Lets not get too technical.  Lets just say that maybe Scott, rather than wasting our time being nice to Bill, should be on the phone with IBM asking whether he can lease a few p5 machines.  I typed that with a straight face.  Honest.

Tuesday, May 17, 2005 (A.G): Jacada Product Revenues Fail To Top $1M!

Jacada, as promised, announced their 1Q2005 results yesterday.  They made sobering reading, and judging from today's downward drift of both JCDA and NETM they definitely spooked some investors.  Jacada, a more credible bellwether of the overall Web-to-host market than NetManage (as intimated on April 26), confirmed, to the chagrin of many, that the market is really as soft and weak as we had suspected (but were hoping, against hope, of being proven wrong).  Revenues were down again, with the crucial product licensing revenues coming in at $800,000!

Yes, there is the inevitable mitigation, and Gideon is quick to spell it out just in case any of us are dense enough not to have worked it out on our own.  Revenues from their traditional business, i.e. Jacada Integrator & Jacada Interface Server(/Anota Terminal Emulator), are going down while the demand for Fusion is still being tested despite their initial boasts as to how Fusion was much sought after by the U.S. Naval Facilities Engineering Command (NAVFAC) and the Latin American telco Embratel.  And then there was the much publicized licensing agreement with Vodafone U.K.

We, however, need to be somewhat savvy when trying to extrapolate these JCDA results to the overall legacy capital market.  Note up top that I was very careful to refer to JCDA as representing Web-to-host.  I did not mention host access.  This was not yet another affectation on my part.  Unlike Attachmate, WRQ, IBM, JCDA does not have a traditional fat-client product or installed base.  They, even after their extravagant $1M acquisition of Anota in 9/2002, are all thin-client.  In this respect they are akin to SEAGULL, but SEAGULL, with their acquisition of Renex, did inherit a large and mature installed base.  This said, we also know that the fat client sector of the market is down too, but if JCDA did have a fat-client base its numbers may not have looked as bad since the fat-client base can always be churned for upgrades and migrations (as demonstrated by NETM).

What these numbers do tell us is that SOA-based host integration is still in the doldrums and Fusion, as a whole, is a compelling technology still desperately seeking a financially justifiable problem.  That, alas, is the problem.

Monday, May 16, 2005 (A.G): Scott McNealy's Modern Day "Munich Agreement"

Scott's new found [i.e. as of last April's $1.95B settlement] desire to collaborate with Microsoft makes me squeamish.  It is one thing not to be at each others throats and resorting to calling Ballmar "Beavis", but it is a whole different thing to try and be so "kissy face" with your once, and still potential, nemesis.  This latest deal to do with single sign-on no less, is way beyond what Novell's Ray Noorda used to refer to as 'co-opetition'. I said it last year, and I am forced to say it again: Scott reminds me of Neville Chamberlain though I do not think Chamberlain was partial, at all, to ice hockey or ever won a tennis tournament when he was young.  This inane desire for appeasement just comes across as a cross between sour grapes and hero worship vis-à-vis Bill.

Scott is reaching out to the wrong camp.  This is why, just last Wednesday, when talking about Gluecode, I said that IBM needs to buy out Scott so that it can control Java.  Scott can no longer be trusted with Java!  Simple as that.

The world and Britain in particular, were fortunate that we had Winston (no not my golden but who he was named after) to bail us out from Neville's folly.  But there is, alas, NO Winston in IT today.  Gerstner, with a cigar, may have come close!  But he is gone.

This single sign-on thing with Microsoft is a travesty.  It is like MS likes to taunt us.  Remember Passport?  Do I have to say more?  Yes, it would make sense to use a Unix box as the basis for Windows security.  But bi-directional?  Give me a break!

Friday, May 13, 2005 (A.G): Is 90% Market Share Something To Worry About?

Another Friday the 13th.  Yet again hardly commemorated despite all the recent interest in the Knights Templar as epitomized in the recent "National Treasure" movie (which to my joy presented much of it quite factually). Anyway somebody had to comment on this being a Friday the 13th, and IF you are at all interested in the background you can find that at http://www.guruge.com/Opinions.htm.

There was an AP report that came out yesterday with the beguiling title: "Microsoft's Share of Browser Market Slips"!  Wow.  I had to go read it.  It sounded significant.

They claim that maybe IE's share of the market has now dropped to 89% from what was 95% last June.  Well I for one wouldn't start dancing in the streets as yet.

As is the annoying wont of so many news stories of this ilk, there is no indication, whatsoever, as to the statistical sample size (or even the methodology) used to make these claims of market share.  As far as I know it was probably, at best, based on this AP reporter talking to 6 of his closets friends.  [But just for 'cics', I checked the stats for this Web site for this year.  On a total of 11,459 hits, 86.4% was IE, 9.7% was Mozilla with Netscape supposedly accounting for 2.7% though I would have thought that that too would have been included in the Mozilla figures.]

I think we all realize, deep down, that this is but totally irrelevant noise.  IE is still top dog and that is not going to change anytime soon.

A few months ago, when Firefox was flavor of the month, I downloaded it and was 'ga-ga' about its tabbed browisng (because I could now have multiple 'home' pages all at once).  I recommended it to family, friends and students.  But then the gloss started to wear off.  IE and Firefox would render the same page differently.  I couldn't deal with that, especially when they were pages (such as these) that I had created.  Since I know that much of the world will still be looking at the Web through IE, I want to see what they see -- rather than a different rendition.

It all boils down to HTML.  It still amazes me that we put up with HTML.  To me all this new effort to refine XHTML is a total waste of time.  In my opinion we need a whole new approach that is not based on markup!  It will only be then that we might be able to break IE's stranglehold of the Web.

Enjoy this Friday.  Have a great weekend.

Thursday, May 12, 2005 (A.G): SOA Maybe the Hottest Topic In IT --
But Can YOU Visualize It?

Just last week (May 2nd to be precise) I expressed my concern that the term SOA is getting bandied around somewhat indiscriminately, kind of like a mantra.  Since then I came across an "SOA in Practice" advertising supplement, in my favorite weekly IT print journal (the only one that I happen to read because 90% of the time they are informative and mildly entertaining).

This SOA supplement confirmed my worst fears as to how this term and concept is now being abused, intentionally or otherwise.  While it is true that I tend to a very pedantic, I hope you agree with me that statements such as "we've created the platform for successful SOA ..." are inappropriate to the point of being blatantly misleading in this context.  Yes, I know what they are trying to say, but I am no longer sure that most of the marketing/sales folks saying these things actually know what they are trying to say!  So I have a very simple proposition.  Get them to draw YOU diagrams of what they mean by their SOA solutions.

Be persistent.  Tell them you want to see some details.

Ask them whether SOA is contingent on a run-time execution model.  If they waffle at that point ask for more diagrams.  This is the only way we are going to break this inane cycle.  As somebody who wrote his 1st book on Web services 3 years ago (and have written another one since), I am very wedded to this whole SOA-centric methodology.  I want, desperately, for it to succeed.  But for it to succeed we need to present the right story and carefully explain what it is all about and how it is going to work.  Just saying 'SOA this', 'SOA that' and 'SOA every other way' isn't going to cut it in the end.  Remember, despite all the hype XML Web services have yet to live up to what was claimed for them.  SOA was meant to serve as a jump start for that effort.  Lets not stall this effort too.

Wednesday, May 11, 2005 (A.G): Despite Gluecode IBM Remains Unglued
When It Comes to Java

Another advocate for open source kindly pointed me at the JBoss Blog for yesterday, which like this one, was devoted to IBM's, rather incongruous, acquisition of Gluecode Software.  From what I can deduce from their tone they are even more put out than I am, though I also get the feeling that they, trying to be ultra-professional, were very circumspect with the barbs they were throwing at IBM -- unlike me, who is much too old and gnarled by now to care whether I offend IBM's delicate sensibilities.  But I still get the distinct impression that the JBoss folks are not amused.

This might be a tad paranoid on their part (and to be fair they are entitled to it when confronting the unprincipled goliath that IBM is of late) but they appear to feel that IBM with Gluecode is trying to undermine their influence and penetration.  That could well be the case, though in the overall Java app. server space, vis-à-vis WebSphere, I have to think that JBoss, despite its immense success, is still akin to tapas on IBM's dinner menu.  They also seem to think that with this move, IBM is making a point to BEA (which could very well be the case) and furthermore sticking it to Sun.

That got my attention.

Sticking it to Sun.

Is that possible?

I have long maintained that IBM and Sun have collectively missed out on Java because of their distrust of each other.

IBM and Sun should have, long time ago, formed a tight, symbiotic Java axis.  But Java is one thing, and Unix servers is another thing.  They can't cooperate on Java because they want to kill each other on the server front.  And Microsoft laughs ...

IBM is yet again devoid of visionary leadership.  Yes, Gerstner should have sorted out the Sun/Java mess before he left -- but he probably, quite correctly, would have thought that he had done enough.  Not sure whether the the anti-trust laws of today will permit it, but IBM, rather than wasting time and money with a insignificantly little acquisition like Gluecode should be working out how they can buyout Sun so as to control Java.  Now that would be sticking it to Sun, BEA and most of all Microsoft.  Think about it.

Tuesday, May 10, 2005 (A.G): IBM's Acquisition Of Gluecode -
The Logic Doesn't Stick

Yes, I fully understand that IBM has an incontrovertible fiduciary obligation to increase shareholder equity.  They, as I have been told numerous times by them, have never, ever claimed that they were a not-for-profit organization.  I, nonetheless, may be dead wrong, but I think many of us in the IT community, somehow would like to think that IBM, unlike say Microsoft, is not totally driven by blind avarice.  Something about social responsibility etc.  Well this Gluecode Software acquisition yet again proves that today’s IBM is a very different, feral, animal to the one that the Watsons envisaged.

Gluecode Software, incorporated in 2003, and with a grand total of 18 employees, has an open-source Java application/portal server based on the Apache Geronimo, Derby and Pluto initiatives.  Think of it as a kind of JBoss look alike.

IBM claims that its primary motive for acquiring Gluecode is to try and garner service/support revenues for this open source application/portal server – which is known as Gluecode Standard Edition.  Think of it as what Red Hat does for Linux.

But I don’t buy it.  As somebody extremely committed to Open Source, I have now maintained for awhile that when it comes to open source IBM at best is disingenuous and at worst borderline duplicitous!  I have said this about IBM’s alleged commitment to Linux and its continued promotion of AIX.  I have said it about z/OS, CICS and DB2 – where IBM will not even open a dialog about the possibility of managed source.  Many of you have probably heard developers open scoff IBM’s contributions to Eclipse.

And now this.

I can, as usual, justify my ranting.  Despite this acquisition IBM still intends to go full-bore with all of its WebSphere Application Server products.  It even plans to convince open source users to migrate to WAS.  This disappoints me.  Talk about tainted money.

Monday, May 9, 2005 (A.G): Will Fusion Keep Jacada Intact?

Jacada will announce their 1Q2005 results next Monday, May 16 -- prior to the market's opening, as opposed to the more usual after market close disclosure (but this may have a lot to do with the 6 hour time difference with Israel).  It will obviously be interesting to see what they have to say since this will give us another insight into the ongoing health of the host access/Web-to-host market to act as a backdrop to the market consolidation that is attracting so much interest of late.

Jacada, along with NetManage, as previously discussed, are the two companies most impacted by the WRQ/Attachmate, SEAGULL/SofTouch, NEON/ClientSoft and Software AG/Sabratec pairings since they are the only two remaining 1st-tier players -- obviously not counting IBM which continues operate in its own space/time continuum.  Five years ago, Jacada and SEAGULL were the up and coming stars of the Web-to-host sector with Jacada commanding a market valuation of $400M when it went public in 1999 though its then revenue stream was only around $15M.  Today's its revenues are up $5M but its market cap is 1/10th what it used to be!  That alone tells you what a rough road this has been.

In case you have forgotten, around that time NetManage coughed up $95M for Wall Data which had revenues in the $100M range.  NetManage, as we saw on April 26, now has a run rate of around $48M, which just happens to be close to 1/2 of what it paid for Wall.  What a market.  Those of us (and yes, I have been around) who were involved with modems in the mid-1980s had a very descriptive term to describe such markets.  I would rather not say what that was in case I too have to claim that this really was a 'wardrobe malfunction'.

Jacada, known for its fondness for technology, might be onto something with its Fusion.  Fusion extends Jacada's reach beyond that of just host applications.  In essence they have extended the transaction capture logic hitherto used with legacy applications to now embrace Web and Windows applications.  It definitely has potential.  With Fusion Jacada may be able to put some distance between itself and the traditional host market.  NetManage allegedly tried to do the same with its acquisition in last September of Librados.  But as their 1Q results indicated this appears to have done very little for their bottom line.  So lets see what JCDA has to tell us about this unforgiving marketplace.