Monday, Feb. 13, 2006 (A.G): IBM's zIIP for z9 - More Hype Than Hardware?

As the article and IBM interview that we have here clearly attest, I am somewhat partial to the Java-grinding capabilities offered by the IBM mainframe zAAP (for z9s, z990s and z890s).  I think we all know that grinding Java, optimally, was not something that mainframes were designed to do and hence it makes sense to have a more bespoke engine, customized via microcode to do so -- especially when such engines do not jack up the MSU rating of a mainframe (and thus the licensing fees for the software running on that machine).  I am also all in favor of the dedicated IFL engines for running z/Linux.

On January 24, 2006, IBM, without undue fanfare, unveiled another specialized engine for mainframes -- the so called System z9 Integrated Information Processor (zIIP), albeit with no availability date other than that it would be sometime in 2006 (which by IBM's books can be 11:59PM, December 31, 2006).  In marked contrast to the zAAP and IFLs, there is something about the zIIP which does not sit well with me, at least for the time being.  To me the zIIP comes across as HOKEY!

zIIP, at least to begin with (when it becomes available), is meant to handle DB2 for z/OS V8 workloads, thereby, according to IBM, freeing up capacity on the other CPs to process other workloads.  IBM ... can you please run that by me again ... very slowly ... and without too many distracting arm waving gestures.

I could be wrong, but for the last 30 years that I was involved with mainframes, I always kind of thought that doing heavy duty database stuff was what mainframes were really, really good at!  So now we need a better engine for doing database stuff?  So what next, a specialized engine for CICS.  Something is not kosher here.

I also have some other reservations.  The zAAP is available on z890s, z990s and z9s (but not on the older z900s and z800s).  The zIIP will only be available on the z9s that were introduced last July.  Yes, I know that IBM has to try and sell new hardware to fund its pension plan etc.  But blowing off customers of relatively new mainframes doesn't sit well with me.  And then there was this claim, in their zIIP announcement, about: "following on the success of the widely accepted IFL and zAAP, the zIIP ..."  All I will say, knowing full well that it will never happen (which is why IBM always pulls this little stunt), is give us some numbers ... PLEASE.  Not sure about IFLs, but I somehow don't think IBM has sold that many zAAPs.

But maybe I am rushing to a judgment here.  Maybe when I have looked at this for longer and from more angles, I may modify my stance somewhat.  But for the time being I can't see myself telling anybody that they should really look at the zIIP.

Sunday, Dec. 25, 2005 (A.G): AttachmateWRQ Wants NetManage

WRQ having conveniently assimilated their prior nemesis, Attachmate, without undue trouble [q.v. my Battle Royal series] now want to acquire NetManage -- logically (and fiscally) the next highest ranking player in this arena [not counuting IBM, given that IBM is likely slightly too "rich" for the investors now driving the so called AttachmateWRQ].  This is not an idle rumor or speculation.  It is based on a document that was filed with the SEC, as in the U.S. Securities and Exchange Commission, on December 6, 2005.  So one has to give some credence to this claim.  As if to validate it, we have not only seen an inexplicable blip in the NETM stock price around that period, but I have seen a marked increase in Google searches for "NetManage WRQ" (ending up at this site) being performed by one or other of the companies.  The most interesting being one performed by a Cupertino employee, on the 23rd, at 6:53am local time!  Wow.  That was keen (and I could even hazard a guess as to who it was).  A quick look at various message boards also shows that this information is fairly well known and that some suspect that a proxy battle for NETM is on the cards.

The link to the document in question was sent to me, unsolicited, by a former NetManage executive.  Given that there has been around 37 executives that have taken a spin through NetManage's ever revolving doors since the late 1990s, with at least 3 VPs (that I know of) bidding adieu this year alone, it is no secret that there is quite an elite club of people out there keeping an eye on what their former company is now up to.  This letter to the NetManage board makes interesting reading.

In particular I was delighted and relived to note that I am no longer the only person that appears to have noticed that NetManage, despite the Librados acquisition, is continuing to lose revenues [q.v. April 26, 2005].  The numbers are not pretty.  I think NETM made $47.7M in 2004.  They are on track to make around $42M in 2005.  This would not have been too bad, if not for the fact that they acquired Librados -- supposedly a major player in the same market in which iWay, the market leader, is racking up around $55M.  You would think that Librados would have brought in at least $5M to help out Zvi.

But declining revenues, however, is the undeniable theme across this sector.  I went and looked at Seagull's 1H2005 numbers that were posted in November.  They were exactly as I had expected.  Revenues are up thanks to the contribution from the SofTouch products. [Comments on licensing revised on January 9, 2006, based on update from Seagull.]  Jacada, at least, is bucking the trend, but their overall revenues are so small (with software licensing at $2M/quarter) that they no longer represent the true market pulse.

The difficulties in maintaining revenues in this sector were predictable and I did talk about it, earlier this year, in terms of the cross-over from "emulation" to EAI.  Hence the vendor consolidation we are seeing.  NETM's current market cap is around $51M.  They claim to have around $22M in "cash".  So if WRQ offers around $55M, it is only really costing them around $33M.  Now here comes an interesting twist.  I had a few calls from Wall Street folks in the Summer re. Attachmate/WRQ.  They kept on talking of debt investing.  Like funding a mortgage -- as opposed to getting one.  In other words, the investors want a 5 - 10 year revenue stream from the investment they are making in WRQ, Attachmate and now possibly NETM.  When you look at it that way, it makes a lot of sense.  There is a revenue stream from the maintenance contacts!  Forget new products.  Makes sense to me.  Against that backdrop, if they do manage to acquire NetManage they could probably get rid of the whole operation and service the installed base using the AttachmateWRQ infrastructure.

If you use this logic, it would also mean that they would also want Seagull.  And who knows.  This may already be in the works.  So by mid-2006, this space could be somewhat sparse with AttachmateWRQ truly ruling to roots and raking in those maintenance revenues.