IBM "manipulated Gartner" - HPQ

But that's not how HPQ sees it, according to this internal document it is sending to all of its partners worldwide. Sheesh!

All italicisations and emboldening are HPQ's. We particularly like the line that says that as eliza is a "vision", it is mostly vapourware. µ

Counter the IBM Migration Sales Pitch
On May 21, 2002 IBM held a 90-minute teleconference to try to convince HP customers (primarily targeting Alpha and Himalaya customers) to migrate to an IBM solution. Overall, this teleconference was a sales pitch loaded with IBM FUD and Gartner inaccuracies. Given that there was not even a single question at the end of the call, it is apparent that all HP customers who listened were able to see through the IBM marketing hype.

IBM cleverly manipulated a generalized Gartner presentation on "customer needs" and "potential migration pitfalls" as "evidence" for why an IBM migration is needed (in reality, no specific IBM advantages were mentioned). This document will alert you to IBM's migration pitch/tactics and provide counterpoints to help you thwart IBM's efforts. Detailed Counterpoints to IBM Migration FUD
Ø Alpha, Open VMS and Himalaya products dead-end after 2 years
o Although the Gartner speaker acknowledged HP roadmaps go out farther , they feel that HP will not honor the roadmaps beyond two years. Since 1995, Gartner has consistently claimed that AlphaServer customers urgently need to migrate within the next two years. They have consistently been wrong and they are wrong on this prediction too. Bring up the fact that this Gartner analyst was paid big money by IBM to say what he said.

o By 2005, IBM will be forcing their own customers to migrate to the Power5 platform . Indeed, IBM's Unix servers are typically replaced by newer systems within 2 years (e.g., F/H/M80, p660-6M1). If Itanium has compelling performance advantages over Power by that time, they may even be moving to Intel's McKinley. These two IBM migrations will be of the same magnitude as Tru64 or OpenVMS to HP-UX on Itanium. Why should customers migrate to IBM today only to migrate again with IBM in 2-3 years?

o The HP support roadmaps clearly state product support until 2011 .

o Gartner incorrectly made it appear as if Open VMS was going to disappear after 2002, when in fact Open VMS will be ported to Itanium (it will be supported far longer)

o Tru64 features (including migration tools) will become part of HP-UX. Customers will not miss the key Tru64 functionality they've grown accustomed too (e.g., TruCluster, Advanced File System (AdvFS), and other features).

Ø Some customers must migrate immediately to respond dynamically to changes in their business
o This is total nonsense! HP will be selling and enhancing Tru64 UNIX, OpenVMS, and AlphaServer through 2006, and supporting through 2011. Customers purchased these systems to support their business needs today. These servers will still be capable of supporting their businesses long beyond 2004 , regardless of the changes in business needs.

o Any customers migrating to IBM's newest servers today will still not have the ability to dynamically partition (static LPARs) or dynamically add CPUs (capacity on demand) - planned Q402 -- and there are no plans to be able to dynamically add hardware online (can do per nPartition with HP-UX). HP has these capabilities today with HP-UX. How will IBM fulfill customers' immediate dynamic requirements?

o Almost all of these AlphaServer customers looked closely at IBM servers at the time of their purchase and rejected IBM either because they were too costly or they did not meet their business needs. Why would they now embrace IBM?

o Many AlphaServer customers were early adopters and risk takers who purchased Alphas because of the leading performance of Alpha or the advanced features of Tru64 UNIX and TruCluster. These are generally not the types of customers who would enbrace an IBM solution.

Ø Itanium is a forced transition that will not suit all customers
o HP has provided superior investment protection by providing dual RISC and IPF upgrade paths even within the same severs. IBM has no such strategy.

o IBM has made many attempts to seed doubt about IPF, but neither IBM nor Gartner had anything specific to say.

o IBM customers know all about forced upgrades as server upgrades are almost always box-swaps (e.g., POWER4 is a box-swap from PowerPC, will POWER5 be another box-swap?) and customers must upgrade both the architecture and OS (POWER4, AIX5L) to get key functionality like partitioning, enhanced workload management, etc.

Ø HP is late with Itanium -- Madison is a "riskier" platform than McKinley because it is farther out
o IBM is trying to imply that HP will leave customers hanging, with no new chip technology until Madison, and this is simply not true. HP has stated plans for PA-8700+, PA-8800 and PA-8900 in addition to IPF. The AlphaServer line goes out to the same timeframe as PA-8900 with EV7 and EV79.

o Power5/6 and most eLiza functionality is way out in the future. How soon can I expect to see this functionality in IBM servers? Will my current server be upgradeable (How "safe" is my current IBM investment given that IBM typically forces box-swaps and major OS upgrades for key functionality/enhancements)?

Ø HP's commitment to cost savings will override commitments to product lines
o HP's overriding concern for the AlphaServer installed base is customer satisfaction, not cost savings. To maintain HP's new #1 standing in UNIX, we must keep the old Compaq installed base happy and help customers move to the best architectures and operating systems. HP is going to work with customers to make this move a smooth one.

o It would be foolish for HP to ignore our product lines since this would only mean we would fall behind in the market and ultimately lose profit. HP roadmaps are flush with planned product enhancements (see http://gtm.corp.hp.com/). What HP is seeking to do is to reduce our own internal costs and overhead so we can keep our prices low for customers, but not at the cost of our products - this is not an either-or.

Ø HP believes in one-size-fits all while IBM provides the best fit for each customer's needs
o IBM's iSeries and mainframe revenue is predominantly (e.g, in the 75%+ range) attributed to installed base customers. IBM is not supporting these platforms to be an additional "choice" to new customers, but as a cash-cow that requires little R&D effort on their part and which they continue to reap profits from costly & proprietary IBM software, maintenance contracts, and services.

o IBM has conflicting strategies - on the one hand they try to justify their multiple server lines by talking about "best fit" while on the other hand they are busy consolidating their product lines by porting the same processors, partitioning, and other technologies to each of them (see http://news.com.com/2100-1001-919579.html?tag=fd_top). In fact, IBM just recently axed its high-end xSeries, which is next to go?

o HP recognizes that one size does not fit all. That's why HP is focusing on UNIX, Linux and Windows for the majority of applications, with specialized operating systems (like OpenVMS and NonStop) for special markets. While IBM has to try to spin their legacy overlapping (proprietary) architectures into some sort of positive message, HP is taking action to consolidate and modernize its operating systems around industry-leading functionality and open standards.

Ø Application availability is an issue for Tru64 and Open VMS customers
o Most Tru64 UNIX customers are running one of the over 6000 64-bit ISV applications supported on Tru64. Of these applications, all but 20 ISVs are supported on HP-UX today. That means a seamless migration on their own timetable for the vast majority of customers. Most customers will wait to migrate until Tru64 UNIX and TruCluster features are built into HP-UX and performance of the Itanium platform exceeds the performance of Alpha. That is expected to occur in the 2005-06 timeframe. In the meantime, Alpha customers can continue to enjoy the features and benfits of the platform without risk.

o Customers with custom applications that only run on Alpha will have an array of porting tools and HP Services available to help with the move. They will also have the Customer Assurance Program if they are not satisfied with the results.

o Finally, don't overlook the Customer Assurance Program (see below). Let HP show AlphaServer customers what we can do for them before they go running off to IBM.

§ A money back customer satisfaction guarantee for transitioning to Itanium processor family-based systems

§ Financial assistance and protection with an attractive AlphaServer trade-in program and a leasing option on current AlphaServer systems that provides a comprehensive transition solution to Itanium processor family-based systems.

§ Transition assurance with free customer architectural transition workshops, availability of comprehensive white papers, toolset, and technical resources and access to worldwide Expertise Centers for testing ISV and custom code on Itanium processor family-based systems.

o Looking at the same Gartner quadrant IBM uses, we see that The ISV support profile for IBM's mainframe Linux is at the very bottom despite IBM positioning this as a "safe" migration choice. IBM AIX is also not broken out by 4.3.3 and AIX5L - these are two very different Operating Systems and AIX5L has only a few hundred applications certified on it and only a handful are 64-bit (basically, few applications can really take advantage of the revised architecture). If application availability is an issue, IBM is certainly not your best bet!

Ø IBM's eServer vision is to unite all IBM servers using eLiza, common middleware and Linux so they fit seamlessly into the customer's application architecture.
o "eLiza" is one of IBM's visions along with "Grid." IBM does not want Sun to be known as the vision company, so they invented their own set of visions. Being a vision, eLiza is mainly vaporware at this time . Many of the same features are being built into HP products and HP has more real "self-monitoring, self-healing" functionality today than IBM with the UDC (per Gartner).

o Another name for IBM's common middleware is "IBM's proprietary middleware." Once you adopt IBM's middleware across your enterprise, you are locked into IBM and that is exactly where they want you to be. HP is pursuing an open middleware strategy. A commitment to HP servers will not lock customers into HP long-term.

o IBM is using Linux to try to unite their many, older proprietary platforms and to try to look more "open" (despite the fact that special Linux versions must be employed for IBM's archtecture - this is not Linux on Intel). It is not for customer benefits, but for IBM's. Previously, they tried to do this with Java but never got traction. In a few years, it will be something else.

Ø What can I do if my customer wants to call in IBM for an evaluation?
o Obviously, customers are free to do what they want to do. Try to find out the driving reasons for the call. Is it to stimulate competition so the customer can get better prices from HP? Is it because the customer is truly unhappy with their existing AlphaServer installation? Is it because the customer is worried for the future of the platform? In all three cases, try to schedule an architectural transition workshop with HP before IBM comes in. This will not only evaluate present and future needs, but will also help explain the tools and assistance available from HP. If a customer is worried about the future, then they are probably reading too much into the competitive FUD. Identify what specific concerns they have and work to overcome them. Call in an executive if necessary. We may also want to schedule our own regional customer concalls to reassure customers.