Friday, June 17, 2005 (A.G): Anyone Still Paying Any Attention to AnyNet

AnyNet, for those of you who may have forgotten, was IBM's rather cute and clever, protocol conversion technology, c. 1992, that facilitated multiprotcol networking by permitting SNA-across-IP, IP-across-SNA, NetBIOS-across-SNA, or even SNA-across-NetBIOS.  Rather than encapsulate, per DLSw or today's EE, AnyNet performed bona fide protocol conversion -- with what was called compensation.  My "Reengineering IBM Networks" book, if you are interested, has more details of AnyNet than any other non-IBM publication.  So check that if you are curious about AnyNet.

Anyway, z/OS 1.7, which will be available in September this year, is the last z/OS release that will include the AnyNet 'gateway' component that works alongside VTAM.  So this is a heads up for those that are still using AnyNet, particularly for SNA-across-IP, where SNA applications on a z/OS is an end-point.  Well this is nothing to panic about. z/OS 1.8 will not be with us till September 2006 and many will not get around to using it till most likely 2008.  But it is worth keeping this in mind since the CICS Transaction Gateway uses this mainframe AnyNet!  That might have got some of your attention.  Have a great weekend.

Thursday, June 16, 2005 (A.G): IBM's Acquisition of Isogon

Isogon, with its strengths in mainframe-centric asset management, indubitably, looks like a fine, upstanding company -- an asset to the mainframe community.  But why bother to acquire it?

Why is IBM expending all this effort, time and money acquiring, piddly little, marginal companies?

Sun is buying StorageTek (June 3) and Apple has ditched IBM.  Against this background, IBM, supposedly laying off people, continues with this inane program of buying, small, insignificant and irrelevant companies.  It was Gluecode a few weeks ago, and now it is Isogon.  What gives here?  To me it is more proof that Palmisano is way, way out of his depth.  BUT, as usual, I can substantiate this claim.

In mid-1994, when IBM was beginning, for the first time, to feel the heat from Cisco, I had a meeting with the then joint chiefs of IBM's Networking Hardware Division (NHD), the semi-legendary "B.J." Johnson and the more mortal Don Haile, in Somer, NY, to propose that IBM acquire CrossComm -- at that juncture #2 to Cisco when it came to SNA/LAN integration, though the gap was rather large.  My friend, Gregory W. Koss, the then Senior V.P. of Marketing for CrossComm accompanied me and as ever gave a virtuoso presentation.  B.J and Don were visibly impressed.  They asked me to stay behind and bid farewell to Greg.  They then told me another Gerstner gem vis-à-vis successful enterprise management (and they had shared a few of these with me, before and hence).

They told me that the effort and the cost to IBM of acquiring any company, irrespective of its size, is roughly the same.  So Gerstner had told them not to bother with companies that were smaller than $2B.  Therefore while they would have loved to have snapped up CrossComm, for around $200M, they couldn't even propose it to Gerstner.  The deal was too small.  As if to underline this point, IBM went onto acquire Lotus for $3.5B a few months later.  I hope you all see my point.

I could drive it home with another cliché from the M&A world: "if you can get all the milk you want why bother to buy a cow?"

Palamasino, you sure are no Gerstner and I am not sure whether you can tell the difference between a cow and bull.

Wednesday, June 15, 2005 (A.G): Super LPAR With 24(+) Processors

Once you start partitioning a data center server, as is now becoming the norm for BIG iron (thanks mainly to IBM), a question that you invariably face is how you optimally balance your processing resources across the LPARs -- though IRD on mainframes makes this a bit easier enabling resources to be automatically switched, 'on-the-fly' between LPARs based on policies and workload demands.  But there is still always the question about how 'big' you can make your 'main' LPAR.  Right now with mainframes, when it comes to processors, the biggest LAPR you can have is a 16-way LPAR; i.e. maximum of 16 business CPs driving a LPAR at any one time.  This was fair since prior to the z990 (in mid-2003) we didn't have mainframes that could have more than 16 CPs in total.

It is since the z990, with its 32 business CPs (though as I talked about on May 24 there might be more unused CPs than we have been told about) that we now have to think about LPARs that may support more than 16 CPs.  With z/OS 1.7, still on track to be generally available in September, 2005, we will be able to have brawnier LPARs with more than 16 LPARs.  But exactly how many is now up for grabs.

z/OS 1.6, the first 'MVS' OS that would only run on z/Architecture machines [i.e. z800, z890, z900 and z990] was previewed in February last year, and was available for GA as of September 24, 2004.  At that time and until recently IBM stood by the claim "z/OS 1.6 is planned to support up to 24 processors in a single z/OS image" -- where this 24 count included both CPs and zAAPs.  The new number appears to be 32, though to get to 32 you would have to have some zAAPs, unless of course you only plan to have one LPAR on a 32-way z/990.  IBM's current documentation is a bit confused on exactly how this 32 can be made up.  But I wanted to give you a heads up since I just found out about it, and as far as I can see there isn't any mention of this increase on IBM's public documentation.

Tuesday, June 14, 2005 (A.G): BEA's AquaLogic Is More Than 'Vaporware'

Since late last week, starting with a very high-profile, gala unveiling at the NYC NASDAQ MarketSite, BEA has been real busy pushing its new AquaLogic concept -- with two big Webcasts scheduled for tomorrow (which some of you may want to catch).  It is thus ironic and rather sad that this major strategy/product family roll-out has been severely dampened by a somewhat harsh (and possibly even erroneous) downgrade of BEAS by Bear Stearns on the grounds that they face stiff competition -- as if that was anything new!

AquaLogic strives to provide a service-centric server infrastructure for managing the lifecycle of SOA-based initiatives.  Unfortunately BEA's definition though close is not as precise -- and therein could be one problem, not to mention the name.  Yes, we all know that "logic", thanks to the once market-leading BEA WebLogic application, is a good BEA 'sub-brand' and that they would like to propagate it.  But it is too hokey, especially when they start talking about liquid resources and the new to thaw corporate assets etc.

It would have been so much better and easier if they had just called it the BEA Composite Application Server -- because that is what it is.  Think of it in those terms and you will see that it all fits together rather nicely.  Actually, on paper, it looks like a damn fine offering (and I know a thing or two about what makes a composite application server).  While checking out AquaLogic, just to confirm a hunch, I also went and did a search on a "whos who" of BEA authorized partners, knowing that WRQ, for one, was on the list.  Yes, it was just as I suspected.  Attachmate, independent of WRQ, is also a partner.  So the list of pertinent partners that I found included: WRQ/Attachmate, SEAGULL, Jacada and iWay.  Wow.  What more can you ask for when it comes to developing composite applications?  This is why I (like a few others) am not convinced that Bear Stearns got this one right.  What do YOU think?

Monday, June 13, 2005 (A.G): IBM Bows To PeopleSoft For Its Recruitment

At this juncture referring to my ongoing concerns, started with my May 5 posting (and addressed again on May 23 and June 6), that IBM is unraveling I could, with utmost justification, just say: QED.

This is straight from IBM's Web site for those looking for a career at IBM: "Thank you for you interest in a career at IBM. By clicking the Continue link below you will be connected to the IBM Opportunity Marketplace Job Search Web site. Please note, this site was developed utilizing PeopleSoft Technology, therefore the look and feel of this site differs from a standard IBM site."

Developed using PeopleSoft Technology ...

The look and feel of this site differs from a standard IBM site ...

Yikes and to think that I, in print [e.g. my Corporate Portals book], more than once, promoted IBM's WebSphere Portal technology.

Yes, I understand.  It is all about maximizing shareholder equity.  Cheaper to outsource with PeopleSoft than doing it with WebSphere. 

I really think I can stop here!  [I, being the cynic, however, do wonder if PeopleSoft skims off all the best people who apply before passing on the rest to IBM.  Laugh.]

Friday, June 10, 2005 (A.G): So Where Will Applimation Fit In?

Six weeks following the momentous, market space warping WRQ-Attachmate merger(/takeover) of April 18th, all still appears to be very much status quo, more or less, at both companies, at least to the outside world, though Attachmate now has a more flashy home page and there is indeed a new management team overseeing both companies -- with Frank Pitt, alak, now history.  Please visit the April 20 and 21 entries for the Frank Pitt eulogies, which from what I can see where, alas, a rarity.  Poor Frank.  All that work in the legacy arena and no legacy to speak of.  It is OK Frank, I will try and keep you in mind for the next couple of weeks at least.

Applimation.

We better check them out.  They claim to be a leading provider of application and information lifecycle management solutions.  And some of you, always skeptical are by now asking "so what?".

Well 2 weeks prior to this takeover, Jeff Hawn, the new head of the combined WRQ/Attachmate conglomerate joined Applimation's board.  Now it isn't as if application and information management is outside the purview of WRQ's interest.  If anything it makes a lot of sense for WRQ to add this capability to their current repertoire.

So ...

Will we see Applimation joining the ranks of WRQ/Attachmate?

I have had some amusing e-mails from WRQ distributors in foreign climes.  They are worried.  They feel that they may get usurped by the Attachmate direct sales folks in those countries.  I will talk about this soon.  Enjoy the weekend.  If this Applimation deal happens ... just remember where you first read about it.

Wednesday, June 8, 2005 (A.G): Anonymous Resolution

When I got an e-mail from IBM inviting me to an online briefing on "DB2 Anonymous Resolution" a few weeks ago, I was at a loss and couldn't for the world of me fathom what the heck they could be talking about.  Though I did not get a chance to sit in on the briefing I did poke around a bit since it intrigued me.  Would you believe that it has to do with compliance again!  This compliance stuff, as we have talked about on May 25 and May 31, really is having a profound impact on IT operations.

It is kind of neat and I can see how it works ... but at the same time I am not 100% sure exactly how this helps, because as the name implies it is all about keeping things "anonymous".  So it is a bit like getting an e-mail (since nobody calls anymore) which informs you that our friend "xyz" is having a torrid affair with our other friend "john doe" -- where the e-mail rather than giving us names, nicknames or even a slight clue, really does resort to totally hiding the identities.  And that my friends is exactly what this technology does.  It allows information to be shared and even correlated BUT hides sensitive information such as names, addresses, phone numbers and most likely e-mail addresses using anonymous digital signatures. It is clever stuff, but what I am not sure, and again as I have said compliance is not my 'thing', is how sharing 'nameless' information is going to help us that much -- especially when all these corporations have already lost all these thousands of records with all of our information in pristine form.  Laugh. 

Tuesday, June 7, 2005 (A.G): Breathing New Life Into COBOL Applications

Having started programming in 1969 (on paper-tape at that), and having been an IT consultant for nearly 20 years, I am the first to admit, without reservations, that I am the cynics cynic -- particularly when it comes to technology that purports to revolutionize legacy applications.  Since 1996, having dealt, hands-on, with at least 60 companies involved one way or another with legacy modernization, I kind of feel entitled to think that I have heard most of it, seen a lot of it and been disappointed by a lot of it.  Therefore, I have to admit that I am in a state of mild shock having had a chance to deal with Dr. Cyrus Montakab of SoftwareMining about his intriguing COBOL to Java or C# conversion technology.

I am sold.  As the Monkey's song goes: "I am a believer." And it isn't just because Cyrus, like me, is a Brit.

This was Serendipity again (and I think I have told you that Sri Lanka, my birthplace, was the genesis for this magical word and that my wife and I both wear rings that have it engraved on them -- a kind of rings-of-Serendipity).  I found SoftwareMining, on this Website, a couple of weeks ago, thanks to Google.  Yet again it made me wonder what life was like "BGE" -- before the Google era.

I started chatting with Cyrus more or less immediately.  The interview posted today was the climax of this interaction.  YOU have to read it.  If you are involved with COBOL applications it will be a technological 'epiphany' -- the only kind that I subscribe to.  Read about the portability joys of Java, though C# might be faster.

THANK YOU Cyrus.  You have given me faith again!  I AM A BELIEVER (and that song always gave me goose bumps).

Monday, June 6, 2005 (A.G): Apple Bids Adieu to IBM

It wasn't a surprise given that the news had been obviously leaked to the WSJ by Apple/Intel over 10 days ago [May 23] (more than likely to avoid been scooped, as in the past, by some blogger).  Nonetheless, now that it is official and irreversible, it still comes as a disappointment to somebody like me who tries, despite the daily setbacks, to think well of IBM (since one way or another it has been the source of my daily bread for 31 years), while at the same time also being a big backer of Steve's amazing vision (both at Apple and Pixar).

Other than Intel, I am not sure whether there are any winners in this.  As I and countless others have said by now, financially this is no big loss for IBM.  But what irks me is that it is such a clear signal of IBM's increasingly poor account control.  Somebody high-up should pay for this with their job and my first choice is Palmisano.  To me he has been a huge disappointment.  I just get this sinking feeling that he is way, way out of his depth.  That is scary.

I am still not convinced with Apple's claims that the PowerPC couldn't deliver on performance.  I think the real issue here is price/performance.  IBM, yet again, was no doubt being greedy and stupid.  Oh, when will they ever learn?  This is not going to be an easy switch for Apple.  Though they obviously prototyped it a long time ago they will now have to rewrite their OS 'kernel' and many of the drivers to support the Intel chipset.  The testing alone is going to be daunting.  But I guess IBM annoyed them enough that they are willing to take that hit.

Well I would like to see Palmisano take a direct hit too.  As with this deal, it won't impact him much financially.  He should be all set on that front.  But as with what it has done for IBM it is all about reputations and respect.