Friday, June 3, 2005 (A.G): Sun's $4.1B Deal For StorageTek

As I have mentioned twice in the last few days alone, viz. May 31 & 25, today's increasingly rigorous IT compliance requirements, with Sarbanes-Oxley Section 404 indubitably in the fore, is a boon for storage vendors.  Remember that sobering statistic that S-O 404 alone will mean that corporations with market caps. greater than $75 million, will by 2006, end up generating, in total, close to 1.6 exabytes of fixed content data per year.  And that is without the never ending growth in e-mail volumes and Web pages.  Consequently the demand for storage will continue to be inexorable and greater than that for processing power.  So you can't blame Scott for wanting a part of this action, though you have to remember that it wasn't as if Sun was not already peddling storage.

So how come Scott isn't being hailed as a hero, especially by the as now long suffering Sun shareholders?  Yes, it could be the 18% markup he agreed to -- particularly since the whole deal is in cash.  That alone is an indictment.  Obviously nobody is expecting SUNW to have much on an upside in the next few years -- even after this deal.  They are probably right and that is sad.

But the real downer here, especially for the SUNW stockholders, is the tacit (but unmistakable) admission by Scott & Co. that they have no choice but to further diversify because their current core business lines [i.e. Unix servers, Solaris and Java] are not living up to expectations.  That is the real scary and distressing part.

To me the ideal deal would have been for IBM to acquire Sun.  Maybe IBM will think about it more seriously now.

Looking through the financial 'press' I happen to see an intriguing little tidbit that there is a $133M 'penalty clause', with the $133M being owed to Sun by STK, if this deal goes awry due to certain specified circumstances.  Wow.  Well the legal fees could end up accounting for 50% of this.  But I have a sneaky feeling that most SUNW shareholders would prefer to get this $67M from STK (in cash?) rather than spending $4.1B of reserves on this acquisition.  Have a great weekend.

Thursday, June 2, 2005 (A.G): Microsoft Office 2006 "Saves" XML

My reaction to the news that the next version of MS Office will use an MS dialect of XML as the default file format for Word, Excel and PPT, is exactly the same as that to an e-mail from Redmond, in April, inviting me to attend the next Mainframe Migration Alliance (MMA) conference as a participant.  Thanks, but no thanks.

OpenOffice, that I talked about on May 6, clearly demonstrates the open source community already has no problems, whatsoever, when it comes to readily reading and interpreting MS Office documents -- without having MS going to the trouble of saving them in XML.  So PLEASE do NOT even think about upgrading to this new version of Office because of this XML capability.  We can do fine without it.  Like I said, thanks, but no thanks.

OpenOffice, for one, already has an XML save mode.  So IF you want your Office documents in XML, (and have wisely not spent money upgrading to Office 2003 which does offer a "save as XML" capability) bring them into OpenOffice and then have it save it in XML for you.  How is that for trying to save you money when it comes to your "save as" needs?

Wednesday, June 1, 2005 (A.G): iWay - The Highway To EAI?

iWay [with 2004 revenues in the $50M mark], a 2001 spin-off from the $300M, enterprise business intelligence giant, Information Builders, is to adapters what IBM is to data center class servers.  Whatever your feelings are about them you cannot argue that the depth and breath of their offerings is unparalleled, and that a true product evaluation would not be complete unless you also checked out what they had to offer.  Consequently if EAI figures in your future, then you have no choice but to become somewhat au fait with iWay.  iWay, from what I have seen, appears to cover all the major bases when it comes to applications and data sources, with their Touchpoint Adapters, that deal with business processes (and associated documents), [e.g. BOMs, purchase orders, and change orders] being particularly germane and compelling.

But I have to give you a 'heads-up'.  They have an unfortunate propensity to exaggerate, which is amusingly strange, since given their unrivaled standing they really don't have to overstate what they have.  Last year, on somebody else's dime, I spent a whole morning diligently tabulating and counting all the adapters that iWay had in their arsenal to verify their then claim of 250(+) adapters.  It was kind of amusing.  Do you want to have a guess at how many adapters they really had?  It is fun.  So start with 250, forgetting the (+), and play around with +/- 10%, +/- 20% etc.

OK.  Ready?

240?  225?  200?

I could only come up with 189 and I had it all categorized in tables so I wouldn't screw-up on my numbers.

So that was disappointing, if not annoying. Why overstate by 30% when you have 4 times more adapters than your nearest competitor to begin with? 

But that said, iWay is a force to be reckoned with and their recently unveiled iWay Software 2005 initiative seems to strengthen their hand even further.  So if you are supposed to be looking at EAI and SOA-based composite applications, spend sometime looking at what iWay is saying (keeping in mind that they probably only have 210 adapters as opposed to 280) -- if you haven't already done so.  Thank you.

Tuesday, May 31, 2005 (A.G): Non-IBM Compliance

Last Wednesday, May 25, I talked about how the Sarbanes-Oxley Section 404 requirements which came into law as of last November seem to have really focused IT attention on the need for compliance adherence though "S-O" [which in reality only impacts relatively large corporations] is by no means the first regulations that necessitated diligent IT record keeping.  But thanks to (if one can really use that word in this sorry context) WorldCom/MCI, Enron et al., corporate executives now seem quite anxious not to run too far a foul on regulatory issues and as such expects IT to do what it takes to be compliant.  The problem being that this is all very new and rather complicated.  So, from what I have heard, nobody is still 100% sure what really has to be done to stay within the law.  Many, wisely, have decided to err on the side of keeping more records than they might need rather than find out, 3 years down the road, that they didn't keep enough.  I think we all kind of know what that is like.  It also reminds me of my days at IBM in the early 1970s when the U.S. anti-trust case against IBM was still in full swing.  I know that we were supposed to keep track of everything -- and in those days, unlike now, we didn't have PCs.

Given the enormous opportunity here to make lots of money, it is to be expected that we will see lots of products to help IT do their job when it comes to compliance.  After having brought IBM's Workplace for Business Controls and Reporting (WBCR) to your attention (without bias), I have also discovered that there is a company in Fremont, CA called Virsa, founded in 1996, whose express mission is to provide software solutions that try and help corporations comply.  As I already confessed last week, compliance is not an area that I know much about.  So as with WBCR, my intent here is just to pass along information that you may find useful.  OK?

Friday, May 27, 2005 (A.G): SaS - The Business End of SOA

I have to admit that I already like this new, soon to be the flavor-of-the-year acronym, much, much better than I like SOA.  SOA, if you have any tendencies towards being a pedant (and I, alas, am beyond cure), tends to be an awkward acronym since it (like what it is trying to describe (and that is a bit of a joke)) is kind of incomplete.  I, at least, have difficulty using it on its own without feeling compelled to embellish it with something along the lines of "SOA-based solution" etc.  But that is me.  But SaS.  I can swing with that with the best of them.

So what is SaS?  No, it is not a misrepresentation of the software company of that name, thought they, I have to assume, must be thrilled with the advent of this acronym.  And it is also not a reference to the Scandinavian airline renowned for its vodka service in business class (a kind of VOS; a vodka-oriented-service).

SaS = Software as a Service.

You have to admit it is more tangible.  Plus we even have, thanks to IBM, an online SaS showcase to prove that we really will see SOA-based solutions come to fruition.  IBM, over the last two weeks, has really been pushing SOA -- like there is no tomorrow.  They even have an IBM, SOA self assessment test that you can take to prove to yourself that you really should talk to IBM about why you need to see a therapist because you can't resist taking silly online evaluation tests that only prove that you should be talking to IBM.  You can find that test at: http://www-306.ibm.com/software/websphere/.  Well it is Friday, and maybe taking this test, assuming that you are already "certified" with IBM [though I think IBM calls it 'registered' to avoid causing confusion] will be a good way to kill 20 minutes while pretending to be doing real work.  Have a good weekend.  Cheers.

Thursday, May 26, 2005 (A.G): Will Nitro PDF Blow Adobe Acrobat  Away?

By now, especially after my rant on April 12, it is no secret that I am not a fan of Adobe Acrobat though given the nature of what I do outside this Website [e.g. write, do research etc.] I am forced to use it to create .pdfs on a daily basis -- sometimes, obviously, with disastrous, "pulling-hair-out-at-the-roots", results.  So much so that I hold Adobe directly responsible for my increasingly rapid loss of hair.  So as I have talked about in this BLOG, I have been on the lookout for a viable replacement of Adobe Acrobat -- though I did pay, full whack, a couple of years ago for Adobe Professional 6.0.

I was thus delighted and intrigued when I heard about Nitro PDF last week; www.nitropdf.com.  I had a look and I knew I had to try it.  I e-mailed them and they were kind enough to send me an evaluation key that enabled me to try it out, in full -- though they also do have a no-charge, "trial-mode" scheme that I probably could have also used, albeit possibly with a "Review Use Only" watermark automatically embedded on my trial pages.

I downloaded it earlier today and played with it for about 20 minutes.  I have to say that I was impressed.  The download, installation, registration and activation was the epitome of smoothness.  So that was a great start.

I then tried it out on a 14 page, graphics intensive White Paper I am working on.  I really was surprised by how fast it was compared to Adobe.  My initial reaction, confirmed with other attempts, is that Nitro PDF is noticeably faster than Acrobat.  Of late I have been thinking that Acrobat is the ideal candidate for desktop grid computing because it seems to be so slow.  By comparison, Nitro PDF was real fast.  So much so that each time I was caught by surprise that it was already done.

So far I have been impressed by three things: its speed, the value-added tool set it offers and its immaculate compatibility.  The later was a joy.  I used both Nitro PDF and Acrobat to open each others documents.  No problems.  But I haven't spent that much time testing to be honest.

There was, however, one thing that I was disappointed with, though it is possible that it is my fault.  I use Acrobat, often, to print or convert Web pages -- often using the Acrobat icon on the IE toolbar.  It is also true that I get a lot of Acrobat crashes as well as misrendered pages doing this.  I kind of now know which pages will work and which will cause Acrobat to go haywire.

So I was anxious to see how Nitro PDF would perform.  There was no toolbar icon so I had to print to Nitro.  It worked BUT the rendering was not what I expected.  The page I converted was my Excite home page.  It didn't look anything like the original!  It also cropped the right hand.  So I tried the same page with Acrobat and was glad it didn't crash.  I then compared them.  Acrobat won that one.  BUT this could be my ignorance of not knowing how to use Nitro.  But rest assured I will put it through its paces over the next few days.  Other than the Web page incident, I really liked everything else I saw.  PLEASE try it.  Let me know what you think.  Thank YOU.

Wednesday, May 25, 2005 (A.G): IBM's WBCR Tries To Help You Comply with IT Compliance

I have been struggling to come up with an appropriate analogy to portray the issue of IT compliance, and I confess that the best that I could do (at relatively short notice) is to claim that it has to be akin to blue whales mating.  Based on what I have seen on TV (and on a recent, hoax e-mail) I have deduced that it is indeed a big deal and involves some sizable equipment.  So far so good, since I know that trying to meet IT compliance now involves having to have very large spreads of storage equipment.  But what really struck me is that unlike with elephants (and this was an analogy I am very familiar with from my days at ITT in the 1980s), it is all done without too much noise.  And to me that was the key to this analogy.  Big deal.  Large equipment.  But very little noise.  And to be honest, I, for one, would feel much happier if we heard more about the rigors (if not horrors) of trying to comply with the increasingly onerous demands of IT related compliance -- now brought to a head by Sarbanes-Oxley Section 404 requirements which came into law as of last November.

Those of you involved with data center storage requirements already know that the compliance driven need for maintaining fixed content data has made this the fastest growing segment of enterprise data storage -- even supplanting that of e-mail.  I read somewhere that thanks to Sarbanes-Oxley, corporations with market caps. greater than $75 million, will by 2006, end up generating, in total, close to 1.6 exabytes of fixed content data per year!  That was sobering since exa bytes means 2 to the 60 -- whereas tera is still 2 to the 40.  Wow.

Suffice to say that IBM, as a supplier of storage, is not unduly distressed about all of these compliance requirements.  But to its credit, it understands the pain that its customers have to be suffering trying not to run foul of all of these new rules and regulations (and I also read somewhere that most IT professionals are still not sure whether their companies are truly compliant or not).  So they have this software offering (also available as an 'ASP'-like hosted service) called IBM Workplace for Business Controls and Reporting (WBCR) that may help you try and get a handle on how well you are complying.  It really is all Greek to me since compliance is not something I have to worry about.  So I am just passing this along because when I did get an e-mail about a new release of WBCR, last week, it did strike me as something that is a real business mover -- as opposed to all this talk of SOA-this and SOA-that.  Laugh.

Tuesday, May 24, 2005 (A.G): The Case Of The Missing z990 Processors

When the z990 [the T-Rex] was announced on May 13, 2003, IBM was very explicit that the z900 MultiChip Module (MCM) supported 12 PUs.  Since I still happen to have all of the announcement material, I can show you this ... like this extract from the z990 Reference Guide (GM13-0229-00):

On the top line, you can see that it says: "Each MCM provides support for 12 PUs ...".  Well I think that this could have said: "Each MCM provides support for 16 PUs."

When I was looking for a z990 MCM picture to accompany IBM's José Castaňo's interview on the zAAP, I came across an IBM "Journal of Research and Development", Volume 48, Number 3/4, 2004, that had a rather detailed write-up of the z990 MCM as well as an intriguing picture.

The first thing that I discovered reading this article was that all the z990 processors are dual-core, à la the POWER4 and POWER5.  That makes sense since one has to assume that there is a lot of commonality between these IBM CMOS processors though IBM keeps on insisting that POWER5s, with suitable micro-code, cannot be used as the basis for zSeries PUs.  This was the first time that I had heard that the z990 chips were dual-core ... and it isn't as if I don't follow this stuff (or have a reputation for missing key points like that).  SMILE.

The article then tells you that the z990 MCM has eight (8) dual-core PUs.  It took me a second to do the sums, but 8x2 is NOT 12 according to my calculator.  It insists, and I even tried it in Excel, that 8x2=16. The picture shown here, with all those cute little boxes, appeared as "figure 2" in this article.

The dual-core PUs are shown in red and numbered PU0 to PU7.  You can count them.  I did.  Multiple times to make sure. No question. 8, dual-core chips. 16 processors.

No!  IBM isn't keeping these as spares -- or if so, not doing so openly.  As you can see from the extract above, out the 12 PUs that IBM admitted to, 2 are kept as spares and another two are used as SAPs.

There could be two easy explanations.  The first is easy.  IBM kept the 4 additional PUs quiet to begin with and plans to dazzle us with a z990 MkII announcement which will boast upgraded MCMs with 16 rather than 12 PUs.  That would be logical.  But there is another possibility that occurs to a cynic like me.

These CMOS processors are NOT as reliable as IBM would like us to believe!  So to be on the safe side IBM is keeping 6 PUs as spares rather than 2.

What do YOU think?  You think I am getting too suspicious of IBM -- after 30 years?  What the heck.  This is but grist for that eternal mill.  Thanks.

Monday, May 23, 2005 (A.G): Intel's Power Play For Apple

More than likely this is but a cute (albeit still somewhat malicious) rumor to get some play on the Intel and Apple stock (at IBM's expense), though one has to admit that there is some merit and logic in Apple considering Intel as a viable alternative to the Power PCs -- particularly now that its OS X is based on a Linux kernel.  If this does indeed come to pass, one could say that this was pay back (big time) for IBM's continued disingenuousness when it comes to Linux and open source software in general (as I talked about on May 10/11 relative to IBM's acquisition of Gluecode Software).

As an avowed fan of POWER processors, where the Power PC is but a watered down version, I would be upset if Apple really does make this switch -- particularly if they use the argument, that I have seen bandied about today in the media, that Intel processors will provide better price/performance.  I can't talk about price given that I am not privy to what IBM (over)charges Apple for Power PCs, but having closely followed POWER4/5 processors for the last 4 years I do know a thing or two about their performance characteristics.  The dual-core POWER (PC) architecture, with its integrated caching technology and super-scaler pipelining, is never going to come up short on the performance front.  While Power PCs do not enjoy Intel's economies of scale, I (having been closely involved with IBM pricing in the past) also cannot believe that realistic pricing (albeit at lower margins) is an insurmountable impediment.  To me this cries out as extremely poor account control by IBM.

I will refer to the complaints I made about today's IBM on May 5, and just say QED!  I have this distinct unease that IBM is out-of-control and unraveling.  It reminds me of IBM in the early 1990s before Gestner started to weave his magic. I really hope this is but a rumor.  In terms of real dollar numbers the Apple deal is still small potatoes to IBM.  You probably will not even see a dent in the bottom line.  But there is all that other important stuff related to prestige, reputation etc.  It will be a huge win for Intel.  So if this happens, we really should start demanding that Palmisano do a William Tell.  I will supply the Apple and maybe we can get a couple of archers from the Linux and Java camps to do the honors.