Sunday, March 16, 2008

ACI Strategy - it's all about choice!

I have just returned from spending a few days in Omaha attending the annual ACE Focus meeting. These two day meetings provide more in-depth technical coverage than is usually found at the regular ACI user events, and ACI customers have been coming for more than a decade to hear the messages directly from company executives.

The picture I have included here is of the venue of the Wednesday night social event – a reception held at a local sports bar called the ICEHOUSE. And I found this extremely ironic as my own involvement with ACI came through my association with the ICE product. For most of the ‘90s, ACI had been the global distributor for ICE and then, as we began the new millennium, ACI purchased Insession, creating a separate business unit that it named Insession Technologies. For nearly six years, as part of ACI it enjoyed a successful partnership with the NonStop community and had provided a number of solutions in communications, web services, and security.

But the decision in late ’06 to fold Insession Technologies into the rest of ACI now looks to be less important for the NonStop community than the complete change in direction we saw in the press at the end of ’07. The press release of December 17 ’07 stated “IBM and ACI Worldwide today announced a significantly expanded strategic alliance to create an end-to-end solution for electronic payments powered by IBM's open technology. As part of the alliance, ACI will optimize a new generation of payment solutions on the IBM System z platform including IBM DB2, WebSphere, and Tivoli software and Crypto-chip technology.”

An IBM executive was then quoted, saying “Payments systems running on IBM System z and ACI payments software address these issues and provide our joint clients with world-class transaction processing performance and the flexibility of SOA through next generation mainframe technology.” Philip Heasley, CEO of ACI then added "IBM's capital, both economic and intellectual, will help us accelerate the availability of our integrated payments framework and make it available on what is already the platform of choice for a majority of the world's banks.”

So what does the future really look like for the large number of Global 1000 companies that depend on BASE24 running on NonStop? Will they be facing a forced migration to the IBM mainframe – a good platform, but rejected time and again by the community for this application? And just as importantly, was ACI really going to stop supporting BASE24 on NonStop? I have to admit, after reading the announcements and talking to users, I was a bit perplexed by it all, and wondered about the future of NonStop.

When I was still working at ACI, the new product, BASE24-eps, written entirely in C++, was becoming available on a number of Unix platforms, including those from IBM, SUN, and HP. Early versions were also beginning to be sold to System z customers. And it is this BASE24-eps product that will find its way back onto NonStop. Customers already running the latest Security and User Interface features have been exposed to some of the components that now make up BASE24-eps and have seen these as just the first steps in the evolution of BASE24 to BASE24-eps.

And that’s when it really hit me - every software company retires older releases of their product. It is just not good business to keep on supporting a product that may be back-level by three or four releases. The sun-setting of BASE24, an older TAL-based product, should not be a surprise for any of us as ACI customers were being encouraged to migrate to newer releases for quite some time. At the ACE Focus meeting, a number of very big users had just migrated to HP Integrity NonStop servers and wouldn’t be moving away from that platform for five, or six, or even seven years. For these customers, their focus was on the migration to BASE24-eps, and on working with the recently formed Migration Team in Omaha. ACI was making it very clear that they would continue to be supported for many years to come following such a migration.

Looking at the bigger picture, I am beginning to see that there’s more than one side to all of this. There’s messages, and then there’s messages! And ACI is performing a pretty miraculous high-wire balancing act. On the one hand, walking hand-in-hand with IBM is conveying one message, as is the sun-setting of BASE24 on NonStop, but then there’s the recent sales successes in EMEA. A few weeks earlier, listening to HP NonStop management presenting new business success stories to the SATUG user community in South Africa, it was obvious that ACI and HP NonStop continue to provide solutions.

Aaron McPherson, an IDC analyst, told an E-Commerce Times reporter back in December, that “the partnership solves a growing problem for ACI, which has adopted an evolutionary strategy that tries to knit together all its payment systems … (yet) NonStop systems are so reliable that it is difficult to get customers to move off of them even if the new generation of software is superior and a lower cost to operate. It's a tough sell to get banks to move because payments tend to be an area where banks are reluctant to make changes because it's one of those sensitive areas." McPherson then explained, according to the E-Commerce Times, “if payments go down because you made a bad decision about your upgrade, that tends to end careers. So banks are skittish about it, especially when you're talking about bringing all payment systems together. That doubles the anxiety."

ACI will no longer be supporting HP-UX. But they will not be supporting Windows either and even Linux, a non-starter across the financial industry, may end up unsupported. The platforms they will be supporting will be System z, System p, SUN and HP NonStop. While ACI executives still reiterate the strategic nature of the partnership with IBM, it will be the customer that controls the final outcome and will be choosing the platform that meets their needs for availability, reliability, and perhaps most importantly, their cost criteria.

Readers of the blog have read a number of references to the plans of Platform Solutions, Inc – a Silicon Valley company with firmware that supports zOS on Itanium. Late last year they took in additional investment funds from a number of companies – including Intel and Microsoft. While PSI and IBM are locked in a legal battle, it wouldn’t surprise me at all to see some movement towards resolution happening soon. In any discussion on choice I find it extremely interesting, should PSI prevail and be allowed to support zOS on Intel Itanium platforms, if one of the options as to where to run BASE24-eps swings back to HP. With its upcoming bladed architecture servers capable of supporting BASE24-eps on zOS, mixed in with NonStop, customers may be presented with choice beyond anything imagined inside of corporate ACI.

Perhaps a little far-fetched? According to the electronic newsletter, IT Jungle, writer Hesh Wiener commented in his September 26, ’06 column “A Joint Assault on the Mainframe Hardware Market” that “Platform Solutions has built a series of large-scale computers that can load and run software written for the System z9 and its antecedents. T3 Technologies will incorporate PSI's technology in a line of midrange IBM-compatible mainframes”. As for the hardware under consideration by T3, then “plans to sell machines bearing its own T3 Liberty brand built using PSI firmware and HP's just-announced Integrity rx6600 servers. These HP boxes have four sockets, each able to take a single-core Itanium 2 or dual-core Itanium 9000. Because T3 systems will use the dual-core Itanium 9000s and will be configured to use at least two cores for systems management functions, each box can have one to six cores running in mainframe mode. The machines will match the performance of smaller IBM z9 BC systems.”

The irony that came with having the social event at the ICEHOUSE sports bar was that even today, the ICE product developed by Insession, has become a crucial middleware component of BASE24-eps. Renamed ICE/XS (for Cross Server support) it is in use on a number of Unix platforms and has become available for NonStop. It may even be used in support of BASE24-eps on the System z. ICE was modeled on early IBM VTAM APPN middleware, designed for the IBM mainframe, and it would be particularly gratifying for a number of us to see ICE/XS return to the zOS environment.

This really does remind me of the scene in Star Wars, as Darth Vader and Obi-Wan Kenobi have their last stand on the Death Star. Darth Vader, turns and says “"I've been waiting for you, Obi-Wan. We meet again, at last. The circle is now complete. When I left you I was but the learner. Now I am the Master." Perhaps a little melodramatic, but in this age, where the mix of hardware, operating systems, and infrastructure are becoming somewhat elastic, and where applications no longer are dependent on technology in a lower layer, nothing can be ruled out.

And this is perhaps the biggest message I have taken away from the ACI announcements of last year. There will be choice. There will be surprises. And while I have not heard any ACI executive support my point of view, I am a firm believer in choice. Everything I now know about the strategy suggests that ACI users will be given a choice – and this cannot be a surprise for anyone. But in the end, and rightly so, it will be the customer that drives the future direction of this historically-strong and very important NonStop partner.

17 comments:

Anonymous said...

Hi Richard,

Loved the blog - you present a very well-balanced view of what ACI may be attempting to do - but what I don't hear ANYONE talking about is GUARDIAN.

Tandem's strategic vision has gathered so much dust I don't think that anyone would even recognize it if those old slides and advertising brochures were brought out and dusted off - yet the advantages of Guardian are as true today as they were then.

Tandem convinced us all that hardware fault tolerance isn't the solution - simply redundant hardware was too expensive - they even said that shared memory wasn't the solution either, nor was a multi-cpu solution that 'voted' on instructions - yet we now see all of these solutions being trotted out again.

Tandem once said that software fault tolerance was dismally behind hardware fault tolerance - and the MTBF for hardware was already at such levels as to render it meaningless in the overall picture of application availability - the capabilities of your application will ultimately drive your numbers up into the rarified air of 5 9's. Of course, any good application is only as good as the operating system under which it runs - and Guardian applications have SUCH a rich environment in which to take advantage.

So who has failed? Have we application programmers failed to extend our applications into ways that realize the dream of the Tandem architecture? Has cost and the (purported) disappearance of (good)Tandem application programmers crushed these dreams?

I listened back in 1980 when Jimmy Treybig used to preach to us from the large screen in a conference room in downtown Los Angeles - surrounded by Tandem professionals. I heard his message - I read the literature - and I've worked on applications that I can't imagine running in any other environment other than Tandem.

So who is left to carry the message?

The original ACI developers are gone - long before HP/UX Tandem has drifted away from touting Guardian's virtues - so why is anyone surprised that HP doesn't know how to market the NSK vision? Evidently in order to sell a Ford you have to stop talking about how cool it is to drive the BMW on the lot.

One of the main reason that I love the GoldenGate software solution - is that it delivers on the Tandem promise for fault-tolerant applications - yet I see so many businesses that implement it incorrectly - as a 'backup replacement'. Can we blame them? There's no one telling them why they should aspire to higher application availability.

I recently had an opportunity to provide a client reference for GoldenGate to a company called DealerTrack. They're running SUN, I'm running Tandem. They're ORACLE, I'm ENSCRIBE. Our applications could not be more different - and yet we share the promise of GoldenGate software. When I began to talk to them about our history they grew quiet - and I must have talked for five minutes straight on how we grew our application from active/passive to active/ standby to eventually active/active. I told them how the active/ active applications are almost alive in how they can be designed to automatically react to a changing environment - and they were honestly excited and could not get enough.

The people are out there. The need is out there. They're not all running Tandem anymore (which is a shame). We still know the difference between high availability applications that are best-in-class, and those applications that are designed with the guiding principle of lowest-common-denominator.

Rich Rosales - BASE24 Development Manager

Marty Turner said...

I couldn't agree more with Rich and Richard!
As a 24 year veteran of the "Tandem" platform I continue to appreciate its value more with each day that I still get to program on it. I've programmed on it as an application customer (as I am now as a member of the PROGNOSIS team and a former member of Insession), as an IOP developer (SNAX/CDF), systems management (DSM – NonStop Net/Master) and as a member of the NonStop Kernel group (SRLs, 64-bit, and Integrity NonStop teams) for almost nine years. I can tell you that inside-out it for the most part it still delivers the promises that JimmyT made years ago (albeit some of them "modernized" with the likes of NEO, NSVA, and NSAA0). None of the HPC (High Performance Computing) mob nor IBM mainframes can deliver the numbers of 9s that the NSK can. Too bad HP is so focused on everything but NonStop to the point they don't realize (or don't really care) what they have bought and could still really sell if they actually tried.
I guess I'll continue to be a hard-core NSK-er no matter what else I work on the balance of my career.

David Kurn said...

When I read what Marty posted, he's simply preaching to the converted. With 18 years at Tandem (become Compaq), I became convinced of the soundness and value of the basic NonStop idea. But like many others, I was concerned over its place in light of the increasing reliability of commodity hardware and its decreasing cost. Even so, Tandem with its unique attributes of software reliability and scalability, had and continues to have its place.

With my recent re-engagement with the Tandem platform, I was both surprised and delighted to see that those factors are still present and valid. But, selling the NSK was never easy. One must educate and convince customers. If only HP realized what a jewel they have here. If only ...

Maybe customers should let HP know what the platform is, because for it to shine again, it needs investment, domestic development and evangelists.

David

Randall Becker said...

I'd like to echo the same sentiments to the choir as well. HP has a real diamond here, if they could only see that the cut glass through which they're looking isn't good enough. Customers continue to seek increasing levels of reliability, but budgets and timing push management to a position of "It's good enough" to use more fragile solutions. It's hard for customers to try to sell Tandem fundamentals to HP Product Management, when there are clearly other priorities for both, particularly since that is backwards. But then, we all struggle with the question of what level of reliability is good enough? Where do we set the bar? I've only had one customer in the past decade who set that bar at 100% and was justifiably upset when any outage was required (you know who you are!). Ironically, more recently customers seem to be setting the bar around 99.9%, which is rather frustrating for me.

Cheers,

Randall

Anonymous said...

The HP marketing team needs to push this message to people that don't already know it... heck, we could probably help them sell it by being the evangalists (if we were in 'sales')...

The best answer to both of you is: Does anyone have Pauline's number? I'd ask Jimmy T, but I'm sure he is comfortable where he is in west Austin, and wants to stay 'retired'! (He does, however, miss the free coffee...)

Bill Honaker
XID, Ltd
NonStop since 1979...
and proud of it!

David Finnie said...

At the risk of further preaching to the converted... the previous posts have nailed it - NonStop fundamentals are great and always have been. HP absolutely has to do more if it wants this important platform to remain a viable choice for customers. And that is exactly what you were saying, Richard - it is down to customer choice. If customers see the platform being neglected, they will exercise that choice.

The life of NonStop can easily be saved, but it needs a real, sustained injection from HP - both on the hardware and software fronts. HP is not a small company - it certainly has resources to throw at this issue. Does the industry want an environment where there are limited options for a "mainframe" solution ? I hope not, and I hope HP sees the importance of the enterprise space. Competition is good - for all parties.

Maybe enterprise systems need a bit of a rethink. Does it always need to be shared nothing vs. shared everything ? Independent processors vs. SMP ? What about a mix ? I recently addended Sun Tech Days in Sydney where a major message was that we are unlikely to see many more large gains in performance of individual CPUs with current chip design. Multi core (and multi-multi core) systems is the way of the future. What about systems with several independent "engines" - each with multiple CPU cores. Each engine could share memory. Then multi-threaded apps (esp. Java) could perform nicely within an engine, but the loosely coupled engines could provide superior fault tolerance. All of this needs rethinking hardware and OS support - and possibly also application design.

HP also needs to really improve the OS services and middle-ware offerings on NonStop - possibly in conjunction with ISVs. Not much has improved significantly over the course of several decades.

You can't blame ISV companies for following the dollars to whatever platforms are attracting them. Hopefully HP can resurrect interest in NonStop - the ISVs would have no choice but to follow...

Anonymous said...

Richard,
The sad truth regarding the "new" ACI product running on an Itanium box with z-OS is, it will not approach the reliability of the same application running on NonStop Guardian. While I agree with some who have posted here that HP does not know how to market NonStop, that is not a problem that is unique to HP. Tandem had the same problem, and Compaq never knew what it had.
I think HP Sr management knows it has something special, and I was encouraged that they tried to repackage it(NEO), but the real sale is the value that is unique to the platform, and the focus needs to be centered there.
Dauber

Palmer King said...

Being an old Tandemite, I too have a technical bias to the NonStop architecture, especially when used correctly; however, it is never wise to make business or economic decisions based soley on religious beliefs.

We have to look at both the vendor and customer sides of the economic equation and understand the forces that keep them in balance.

Customers have always struggled with the conflicting forces of buying "the best bang for the buck" historically in conflict with vendor independence. But they didn't weild the clout to force the issue. Pick your vendor and takes your chances.

Not surprisingly, for the software vendor, the same is true. How does one support multiple platforms with minimal code differences?

The hardware vendor, until recently, struggled with keeping differentiators, reasons for the customer and software vendor to prefer their platform.

Then came massive consolidation as the larger companies swallowed up the smaller ones, and differentiators became a burden to the hardware vendor as well. For example, two of the "big dogs", IBM and HP face this dilemna.

IBM now has to support MQ across platforms. Unless they hire a cast of thousands just for MQ, the only solution is POSIX compliance or some semblance thereof. It is just too expensive to do it any other way.

HP absorbed Compaq which had absorbed Tandem and DEC, and are now stuck supporting all the platforms they absorbed as well as HP-UX, Linux etc. Additionally, HP still sells PCs with Windows. Their recent decision to focus on Eclipse and Open Systems with .net as a necessary evil(?) "gives us pause"?

Summarily, for the first time, cross-platform compatibility and the necessary infra-structure to support it is now being driven by the vendors, not the customer.

This is nothing new. It happened with telephones, radio, television, ciruit boards, yada, yada, blah, blah. Why are we so surprised that engineering disciplines and the derived benefits are finally returning to software?

Anonymous said...

Richard,



I read your article (very interesting) as well as the readers comments.



Here are some thoughts from my end:



I think discussing the "platform choice" is a bit like discussing religion or politics - everyone is entrenched in his personal viewpoint and there won't be many changes. As many reader comments stated, we are preaching to the choir



You can do “anything” on “any” platform - the trick question is "which platform do I choose given my requirements (cost, availability, security, performance, ...)". As much as we "Tandem guys" love the platform, it is “not” always the answer !).



In fact, NonStop is the answer in rare cases only and I do doubt how much additional selling the platform can do as long as selling is based on achieving five or more 9s. That is a very limited space … but then, has there been any consideration given, by the user community, to ask HP to provide a presentation (perhaps at HPTF?) where they would (a) bring in HP tech sales people for NonStop, HP/UX, VMS; (b) pick a small number of "application scenarios"; and (c) have each argue where "his" platform would be best? Perhaps outside the scope of what HP would like to see, but from a user perspective, could be very enlightening.



Thomas Burg, CISSP

CTO
comForte GmbH

Richard Buckle said...

Community,

I am really impressed with the scope and quality of the feedback here - enough that I will work this into a blog posting later this week as I want to make sure none of it's lost to the borader HP user community.

Again, many thanks for these comments and I am sure there will be a couple more ....

Anonymous said...

Hi Richard,

I am as impressed by the level of feedback here as I am by your ability to write these blog postings with as much enthusiasm today as the day you started. It's something you enjoy, find gratifying and it's catching on.

The same can be said for those NonStop users out there (some of them posted comments here). Working on that platform has been a gratifying experience and they can really see the value in the platform, even though not everyone knows about the technology.

As a software vendor of Security solutions for the NonStop, I don't mind saying I have been surprised to learn how advanced the NonStop security offerins are in comparison to other "Mission Critical " platforms. The number of times XYPRO has been asked if we can make our products work on Unix because "they don't have anything like that on those boxes"..... was enlightening, to say the least.

There are many facets to determining the applicability of a platform and then determining which one is the best for a given situation. The NonStop isn't lacking in either department and often outshines all other options.

If only there was some way to get the word out. ;)

Lisa Partridge
XYPRO

Anonymous said...

Have you ever heard the expression "Good enough is the enemy of Best"? I think I first read it in a Jerry Pournelle column years ago. It is probably much older than that.

I think what has happened is that as the general public has come to expect, tolerate, and accept that computer systems will get screwed up and stop working from time to time, businesses see that they don't have to spend as much to keep their systems up all the time. That isn't universal, of course, at least not yet, but I believe it is an important factor in shrinking the market for NonStop.

You might describe it as a race to the bottom in quality of computers, just as we've experienced a race to the bottom in quality of so many consumer goods.

Anonymous said...

Hi Richards
Very interesting article. However, from the region I operate (AP) there is quite a stir post ACI's announcement of aligning with IBM. And current customers of BASE24 as also those evaluating EFT Switches are getting mixed messages.
I also know of solution providers undertaking a port of their EFT Switch onto the NonStop Integrity to take advantage of the great features of the NonStop platform.
As you rightly said this will indeed give customers choice.

Mark Green

Richard Buckle said...

Mark,

Good observation (about AsiaPac/Japan market) - perhaps you are aware too that recently in the past 18 months, one IBM mainframe user who converted to HP included a NonStop in the mix (with ACI BASE24) for the transaction component. Not sure if this is also a reference to the same EFT Switch company that you mention, but it wouldn't surprise me.

But the point is, the change in strategy by ACI and the decision to have IBM as their strategic partner, has certainly created a lot of discussion within the community and the results haven't been all that bad so far. It will be very interesting to see how this plays out.

For those that read TandemWorld newsletter, take a look at the piece I wrote in the May 2008 issue:
http://www.tandemworld.net/newsletter%20may08.htm

The Customer is King!

Thanks Mark

Anonymous said...

Richard
The TandemWorld newsletter article is very interesting. Major decisions and purusant announcements of strategy, such as that by ACI, are bound to cause churn in the market. And open opportunities to other vendors to win some of the customers over. S1/Postilion, Distra, Alaric, Openway seem to be making aggressive moves to take advantage of this situation. I see heightened activity by these organizations in the AP market.

Anonymous said...

Hi Richard
On reviewing your blog and the comments I realize that I didn't add my name at the end of my comment that I posted on 19th.

Apologies

Mark Green

Anonymous said...

Hi Richard
Just came across this news item announcing availability of a new switch on NonStop platform:
http://www.sda-asia.com/sda/news/
psecom,id,19919,srn,4,
nodeid,4,_language,Singapore.html#

Mark Green