__/ [ Rex Ballard ] on Saturday 18 March 2006 16:34 \__
> Roy Schestowitz wrote:
>> ,----[ Quote ]
>> | The company (IBM) has for years marketed its products to the tech
>> | elite within big companies. Now Microsoft is making concerted effort
>> | to speak the language of top executives.
> And since the introduction of PWC, now known as BCS, IBM has been going
> right for those same top level executives. Today, the IBM engagement
> process usually starts by focusing on the customer's business problems.
> Things that will help the business grow by increasing profits. This
> means making sure that all IT investments provide more productivity at
> lower costs - based on not just IT costs, but the costs of the
> businesses as a whole.
> The irony is that IBM is now providing Linux support and services
> because the customer is far more interested in the services Linux
> provides, than the costs of the licenses. The combination of OSS
> software for Linux, along with high-quality JAVA based products has
> turned out to be a powerful combination. Customers are far more
> concerned with issues like how to make sure that they have an accurate
> and up-to-date picture of the company's financial status - than how
> pretty the GUI screen on the administrative console looks. They are
> far more interested in workflow solutions which can make sure high
> priority tasks get done across the organization - than in dancing
> paperclips. They are more interested in being able to have a team of
> 50-100 people from all over the world, including New York, California,
> China, India, and London - working together as a coherant team, than
> they are in whether their employees can play "Flight Simulator" on
> their desktops.
> Linux and Java provide these capabiltiies in a way that works very
> efficiently with the rest of the IT infrastructure, especially the
> large enterprise level applications such as SAP, Peoplesoft, JDE,
> Seibel, and other ERP, CRM, and Application Server products. Linux is
> also fully compatible with the infrastructures used throughout the rest
> of the interprise, including security, communications and connectivity,
> and even programming environments.
> There are very few companies these days who don't have at least some of
> these applications running on some flavor UNIX or Linux. And Linux
> provides the ability to provide the same security, stability,
> scalability, compatibility, and supportability as UNIX from the desktop
> to the "Mainframe".
> Linux "Appliance" systems such as Linksys routers sell for $40 each
> these days, and Linux servers for prototyping and pilot projects can be
> made from "obsolete" servers and desktop machines - then rolled out to
> blade servers, scalable Linux servers, Linux on Mainframes, or even
> Solaris, AIX, or HP_UX servers, thanks to the portability of GNU tools
> and Java.
>> | Why? Because business managers, marketing executives and other
>> | nontechies are increasingly involved in technology purchasing decisions,
>> | Microsoft argues.
> Ballmer: Let me be more clear: We're not anticonsulting. I just came
> off the board of Accenture recently. I've been on the board four or
> five years. This is not about being against consulting. This is about
> being for empowering people, so they can they can empower their
What utter rubbish, as well as hypocrisy.
> Ther is a huge irony here. It sounds to me like Ballmer is getting
> that massive "Wake-up Call", similar to the one IBM got in 1991. He's
> been pitching the daylights out of Vista and corporate customers aren't
> willing to sign up for in just because Microsoft wants to sell it.
> This is a very new experience for Microsoft, much like it was a new
> experience for IBM in 1991, when it was UNIX that was pulling the wind
> out of the sales of IBM's ES/9000 and MVS 4.0 products. Customers
> weren't willing to fork over almost double the value of their current
> depreciated IT inventory for upgrades which didn't offer real
> productivity gains.
Well, then come to consider the fact that as of last year (citation on
demand), half of all businesses in America still used Windows 2000. If IBM
cannot sell Windows, push IBM aside and take its role. *That* appear to be
> When IBM tried to introduce MVS 4.0 it solved a big problem - it
> provided a Linear address space, eliminating "the line" - applications
> had to be coded to run above or below a physical memory limit (I
> believe it was 16 megabytes) whach was beginning to become a big
> problem as applications became more real-time oriented. The problem
> was that IBM was still pushing proprietary protocols (APPC/LU6.2/SNA,
> 3270/LU2/PU2.0 token ring) and was doing everything it could to exclude
> UNIX. The problem was that by 1990, UNIX was becoming a common fixture
> in most corporate environments and many PCs were already talking TCP/IP
> as a way to talk to UNIX computers. Many PCs even used UNIX systems
> such as the MITEK gateway to talk to the mainframes.
> When IBM introduced MVS 4.0, it was necessary to replace CPU cards,
> upgrade CICS, IMS, and most of the applications running on the machine.
> Many of the machines being upgraded cost about $4 million each when
> they were new and the upgrades were going to cost $5 million to $6
> million per machine to upgrade machines that were already 2-3 years
> old. The alternative, replacing the machines completely, meant taking
> a big hit on the leasing or financing and paying early cancellation
> penalties - effectively raising the cost of replacements to between $8
> million and $10 million per machine.
> IT managers quickly began exploring ways to migrate some of their
> applications to UNIX and eliminatet the need for these expensive
> upgrades, or at least postpone them for a year or two. The problem for
> IBM was that they suddenly found their quarterly revenues were sharply
> down, with no way to build them back up. Worse yet, Microsoft had
> pretty much killed the market for OS/2 with Windows, largely because
> IBM would not disclose the 32 bit hardware architecture for
> Microchannel to the other OEMs, even for license fees. Internally,
> they were already working on 32 bit OS/2 but the OEMs no longer trusted
> IBM to give them a square deal with OS/2. As a result, the OEMs
> switched to EISA, VESA, and Windows. Killing another critical source
> of revenue.
> By 1992, IBM fired John Akers and replaced him with Lou Gerstner, an
> outsider who wanted the job, not because he loved IBM's products and
> wanted to "sing the company song" but because he knew that if he could
> focus IBM on giving the customer what the customer wanted, he could
> turn the company around.
> Ironically, when he changed the commission structures and bonus plans
> to one in which the customer could choose any option and commissions
> wouldn't be pushing the sales people to push MVS or OS/400 over UNIX
> systems, the customers were demanding UNIX, and wanted to see RS/6000
> systems. IBM introduced the SP/2 line, which was an arrangement of
> RS/6000 based UNIX systems designed to provide mainframe like
> performance for specific functions such as running relational
> IBM also started encouraging divisions and sales organizations to
> support customer preferences. For example, mainframes were desgined to
> support oracle, and software applications such as DB2 and HFS were
> redesigned to run on SUN machines as well as IBM Unix machines.
> The biggest change, the introduction of IBM's consulting organisation,
> now known as IBM Global Services or IGS. The interesting thing about
> IGS was that the staffing was deliberatly "platform agnostic", and they
> looked for people with a breadth of knowledge across multiple platforms
> and multiple product lines. They looked for DBAs who had worked with
> Oracle, DB2, Sybase, and SQL Server. For Unix administrators who knew
> AIX, Solaris, AND HP_UX. For programmers who could work with Windows,
> UNIX, and MVS nd CICS.
> The IGS consultants were deliberatly kept separate from the software
> group support teams, so that the IGS team could focus ONLY on solving
> the customer's business problems using the technology the customer
> already had in place. When the customer needed new technology, IGS
> could offer IBM products, and at that point, the SWG teams could
> provide the product-specific support such as installation,
> configuration, and specialized consulting. It was IGS who led the
> engagement. IGS tried to focus on IT Architects and Specialists with
> broad experience.
> Here comes
> And I think IBM's point of view is it's all about the consulting
> project or the outsourcing. What we would tell you is...we believe in
> giving people the tools that let themselves collaborate, get insight,
> deal with exceptions. One of the big differences between good companies
> and great companies is that great companies deal with the problems and
> the exceptions a lot better than the good companies.
> IBM provides as much or as little as the customer wants. One of the
> big challenges these days is that IT requirements and technologies are
> changing very rapidly. Many people don't have the time or the budgets
> to retrain hundreds of IT workers, many of whom have no interest in
> learning new technologies on their own.
> The demands change radically, even from year to year. From 1995 to
> 2000, a huge percentage of the IT workforce was focused on Y2K issues,
> which meant programming in COBOL for CICS, IMS, and MVS, and Visual
> Basic for Windows. Then, by March or April of 2000, there was no more
> need for Y2K people. Suddenly the top priority was getting web servers
> connected to Inventory, customer databases, and ordering systems, which
> meant programming in UNIX, PERL, Java, Apache, and HTML, MQseries, and
> Then came Sarbanes-Oxley, HIPPA, and the need for business-to-business
> interconnections - not via keyboard/mouse intensive web interfaces, but
> via XML, SOAP, and JAVA. But this had to be done in a secure way,
> eliminating the risk of intruders in one company accessing the
> information of a competitor's through a customer site. IBM called it
> "Web Services". Sun, IBM, and BEA provided tools to do this type of
> integration. The problem was that when Microsoft tried to join the
> came with .NET, the security systems were incompatible, there were
> proprietary extensions, and there were serious incompatibility
> Today, the big business drivers involve workflow, not just passing
> e-mail around, but coordination of updates that can often involve
> dozens of different companies and dozens of different applications and
> systems. Even the automated processes often require realtime
> notifications and reporting - for Sarbanes-Oxley compliance. CEOs
> can't even speak in an interview anymore unless they have a very
> accurate picture of the company's exact financial status - often on an
> hourly basis. They need to know exactly what their sales pipelines
> look like, what their backlogs look like, what's getting returned,
> what's not selling, what's selling faster than expected - so that
> decisions can be made by all levels of management on an hour-by-hour
> Today, the key requirement is compatibility. Compatibility between
> systems that run Solaris, AIX, HP_UX, Linux, FreeBSD, and even SCO
> Unix. And Microsoft has been doing everything in its power to fight
> and resist that compatibility. They keep hoping that if they add
> proprietary extensions, that they will be able to get people to replace
> their Mainframes with Windows servers.
The initiative to promote Windows servers, using those infamous FUD
campaigns, is most likely intended to inject incompatibilities at the
network level. It was only a few weeks ago that I was 'sent' a contact using
some obscure protocol that is associated with Exchange servers (apparently a
request to 'pull' contacts off the mail server). Needless to mention, I had
to demand something standardised. The administrative staff does not
understand what it Outlook-compatible and what it standard.
> IBM has had to teach elephants to dance. They had to make MVS more
> like UNIX to remain relevant in the marketplace and these days, many
> customers put Linux VMs on Z-Series machines to minimize the need to
> integrate APPC, LU6.2 and other "legacy" applications and protocols to
> systems outside the mainframe.
> IBM has had to beef up the UNIX lines, offereing P-Series machines such
> as the P-590 which is nearly 1000 times more powerful than the old 3090
> machines, and about 1/4 the size.
> IBM has created storage systems that can be shared with AIX and Linux
> allowing terabytes (1 thousand gigabytes) , or even petabytes (1
> million gigabytes) to be accessed with response times in tens of
> milliseconds, faster than "the blink of an eye". Some of these systems
> can serve up over a million transactions per second, at less than $6
> per trasaction per second (TPM-C).
> But the real challenge is that these new *nix centric systems can
> provide "up-to-the-second" charting if desired - showing the state of
> the business in much the same way that network managers can view
> "up-to-the-second" status of critical network components.
> I can send out an RFP document in Microsoft word format, get back a
> response, cut-paste all of the numbers back into enterprise systems,
> and after a delay of 2-4 weeks get the right information to the right
> people who can make a decision.
> Or I can send out the RFP document in OpenOffice format, complete with
> spreadsheets, and forms ready to be filled in, and when the response
> comes back, I can use a SAX parser to scan that XML document for the
> requested information and have it moving up the chain of command in a
> 2-3 minutes, with the information required for a final decision in a
> few hours.
You are assuming that OpenOffice formats are prevalent. This is not the case,
yet. With the development and quick penetration of OpenOffice, peripheral
extensions will follow, which will be re-usable. Firefox benefits from
similar contributions that come from the users community.
> The doctor can get those HIPPA records in 2-3 minutes, from anywhere in
> the world, and know immediately of allegies, medical conditions, and
> situations that might affect treatment - before they accidentally kill
> the patient with the weak heart, the allergy to antibiotics, or the
> sensitivity to eppenephrine.
> The retailer who sees a run on a particular style, or gets a bunch of
> requests for a dress in size 14 that they currently only have in Size 8
> can get the requested dress, in the requested size in 18 to 48 HOURS
> instead of 6 to 8 weeks. They can even schedule an appointment with
> the customer so that they can try it on the day it arrives.
> It's nice to see Ballmer singing this song, but he doesn't seem to
> understand that in order to play in this new market, he has to play by
> that market's rules. Rolling out Vista with a bunch of incompatible
> "lock-out" strategies desgined to prevent customers from having full
> access to UNIX capabilities and compatibilities - including Linux on
> the desktop - is more likely to lock Microsoft OUT of the very market
> they are trying so desparately to enter.
Intersting point. I notices that more and more people around here, whose
desktop urges them to make use of lockin, are being told off, by people like
myself. If the degree of lockin is increased, that message will become
louder. In due time, proprietary software will vanish and give way to
> Steve Ballmer wants corporations to stop going to companies like IBM
> who can provide services ranging from "ad-hoc" short term customization
> engagements all the way up to hosting entire enterprise systems the way
> some companies host web sites, all fully integrated into the existing
> corporate enterprise architecture. Or companies like Novell, Red Hat,
> and VA Linux who can manage the hardware sitting on your premises using
> remote VPNs. About the only time someone actually has to go touch the
> machine is if there has been an extended power outage that completely
> drained the UPS.
> Instead, he wants them to buy thousands of little Windows servers and
> Vista workstations which can be managed by high school students - since
> their primary duties will be "rebooting the boxes" under the direction
> of one or two people whose job it is to figure out which boxes need to
> be rebooted.
It becomes harder. The BSoD is sometimes forsaken in favour of La-la land
with some grey prompt.
> Ballmer's approach might make sense for a 4-man consulting firm who
> actually show up in an office every day. They only need one server,
> they don't have too many customers, and the cost of a failing server or
> a virus is probably only a few thousand dollars in recovery time.
> These are the references Microsoft likes to use for their "Fast Facts"
> comparisons to Linux related to TCO. Yes, for those guys, it might
> make sense. They've never used anything but Windows, they don't even
> want to learn Linux, they are 3-5 years from retirement, and most of
> their consulting work involves creating reccomendation documents which
> get written in Word/Excel/Powerpoint, with maybe a few Vieo diagrams,
> and MS-Project Schedules (pasted into powerpoint or Excel for the
> document) and then printed up and mailed in hard-copy to clients who
> will probably implement the reccomended strategy (for which they paid
> thousands of dollars) using Linux, UNIX, and Java.
> The irony is that they might want these documents in electronic format,
> and they will have to be converted to PDF format to make sure that they
> are not revised from their original versions.
Adobe are selling software for PDF export (bound to Office). The University
pays a lot of money for this. I believe that the next version of Microsoft
Office will catch up with OpenOffice in the sense that it will enable PDF
export 'out of the box'. This shall be intersting.
Roy S. Schestowitz | "How do I set my laser printer on stun?"
http://Schestowitz.com | SuSE Linux ¦ PGP-Key: 0x74572E8E
4:30am up 10 days 21:07, 9 users, load average: 0.93, 0.72, 0.88
http://iuron.com - help build a non-profit search engine