System iNetwork

IBM has to do the right thing for all involved and either sell i5/OS or open source it. Letting it run at AIX + DB2 prices on p5 hardware would help a lot, but the OS still needs core improvements such as long object names, native ascii/unicode storing of character data and 32 byte pointers necessary to break the 16meg segment limit. IBM may cut the price, but it will not invest in the product to add these needed features.

Is there any thought or mention within IBM of selling i5/OS?

-Steve

With a "Grain of Salt": these IBM figures are in context of some terrible financial news coming from Wall Street.

The Sub Prime mortgage disaster has metastasized way beyond the Sub Primes into what were considered healthy mortgage and other debt instruments. Now the meltdown is affecting consumer spending.

The Real Estate slowdown is expected to last, at least, until 2010.

--John

The amazing thing about those numbers is that they aren't anything close to what we see for configuration requests on System i.

We see tremendous volume in that space, and our teams are running at full tilt, but for some reason it isn't translating into sales.

The System i revenues may be down, but it is probably still a goose that lays golden eggs. Just because the goose is laying one egg per day instead of two, there is no reason to BBQ the goose.

Also, we need to look at the role of System i in revenues generating for IBM in a broad sense. Some companies practically give away hardware so that they can make money from software and services. For example, most of us (if we are not too fussy about fashion trends) can very easily get cell phones for free with service subscription contracts. One may say that the phone company is losing money on the cell phones. Nevertheless, the company as a whole is making profit because of its marketing strategy of giving away free cell phones.

Despite all the doomsday’s scenarios, System i will likely still have a major role to play in IBM's revenue generating process for sometime to come. Of course, we are hoping that System i's direct role in IBM revenues generating process will increase and not decrease over time. It is good to be a supporting actor/actress but it would be even better as the lead actor/actress.

The 20 percent decline really is a non event. System i revenue declines on standard average by 10 to 20 percent each quarter over quarter since the peak in 1999, with the exception of a slight reversal in 2005 because of POWER5, so no big deal.

What I do appreciate from IBM is that they still report System i seperately, albeiit probably in smaller print or as a tucked away sidenote. Under the new reorganization, they're not obliged to breakdown the figures by hardware productline anymore.

As far as disk sales in the Midrange area, no wonder they're in the crapper if Rochester keeps asking 5 to 10 times more money for the same disk on System i compared to System x or p. But that's probably to complicated for them to figure out.

"...The System i revenues may be down, but it is probably still a goose that lays golden eggs. Just because the goose is laying one egg per day instead of two, there is no reason to BBQ the goose. ..."

Would be useful to know the actual sales and profits of the i5. On Monday, ITJungle reported a sales estimate: 2007 3Q $237 million. That would put us at $1 billion per year. Approx 1% of IBM's total business. Put another way, the top level execs at IBM can spend 2 to 3 days a year ( 1% of their time ) on i5 related issues. That is one little goose!

-Steve

IMHO, I think the direct revenue from System i is only the tip of the iceberg of the financial benefit that System i is contributing to IBM. Focusing only on the direct revenue of System i trivializes, belittles, and falsely simplifies the role of System i in the IBM revenue generating process.

IBM must be leveraging System i for its software (like Websphere, Lotus, Rational) and its consulting services (e.g., Global Services).

The pressure to reduce hardware cost will always be there and we have known that for years -- the hardware cost will continue to decline whereas the manpower cost will continue to increase. Thus, it makes sense to provide low-cost hardware in return for expensive services (as in the cell-phone case that I mentioned earlier).

In the IT Jungle, it was quoted that for System i: "Shipments were up in the double digits, in fact, and I would take a stab in the dark and say that even though revenues have been down, aggregate CPWs of processing power shipped has been flat or rising over all of these years of revenue decline." (some what taken out of context but the idea is there)

Of course, we hope that the tip of the iceberg will grow bigger, which implies that the whole iceberg is getting bigger too.

"...IBM must be leveraging System i for its software (like Websphere, Lotus, Rational) and its consulting services (e.g., Global Services). ..."

Chris can correct, but the numbers I have read are that system p sales are 4x that of the system i. Same with system x. The IBM software and global services mentioned above involve all the IBM hardware lines. By my calculation of system x being 4x the i and system p also being 4x the gross sales of the i, IBM is selling 8x the amount of software and services on the x and p combined compared to the system i.

-Steve

The revenue will depend on the margin. To boost sales, the margin for System i may be lower than System x and System p.

The more units of System i are sold, the more the leveraging power of System i.

In any case, I guess there is no point quibbling over this. We want System i to do well. Let's focus on providing comments and suggestions to help enhance the sales and margin of System i for the good of the ecosystem.

Cheers

What is System i revenue? Does it include revenue from i5/OS or other software running on the system? If not, then what we may be seeing is simply the price of the hardware dropping.

At one of our user group meetings a representative from IBM indicated that System i shipments tend to increase each year, and the CPW ratings increase each year, but the price keeps dropping disproportionately, which is a good thing for customers and ISVs, though IBM needs to make up the difference through additional software and service revenue - which is what appears to be happening.

Since System i and p are virtually identical hardware-wise and neither i or p business units exist anymore, I suspect that at some point IBM will stop reporting them separately, and simply report Power system sales.

As the System i supports integrated System x servers, LPARs, Linux/AIX partitions, AIX runtimes via PASE, Java Virtual Machines, PHP, mySQL, etc., it becomes less and less distinguishable from Intel and System p hardware.

Hardware is becoming a commodity, and even System i hardware is gradually adopting commodity pricing.

[Note from Chris: Hey Nathan . . . definitely the lowering cost of hardware is a factor, and it's a factor across most hardware platforms. I think the concern with System i revenue dropping is that it has been dropping at rates exceeding drops in both p and z, and comparatively, has often dropped substantially while p and z have risen substantially. Of course, System i sales are also cyclical, based on the latest and greatest releases from IBM, but still, the trend of consistently reduced revenue, compared to previous years, is troubling. No doubt, the System i's total cost of ownership still rocks compared to every other platform -- even more so as hardware costs decline. The challenge for a hardware company, the last few years, has been to sell to enough new customers to offset the market-driven decline in hardware/processing power. It's that hard-to-know rate of new customer sales that's concerning to industry watchers who don't get to see IBM's internal balance sheets.]

Yes Chris, the continuous revenue drops must be more than price. We see database and transaction processing applications migrating off the System i and over to Wintel platforms in every industry.

The irony is that for all IBM has been doing to make the System i more like other platforms, the more it has hurt the platform. IBM is following a server homogenization strategy that blurs the distinctiveness of i5/OS, takes away from its traditional value proposition, and alienates the traditional user base, but so far hasn't attracted many newcomers. I guess time will tell whether the homogenization strategy pays off.

I don't think it will, though. It appears to me that the native virtual machine is so much superior to the other virtual machines (Java, PASE, PHP, LPARs, etc.) that only a return to it will turn the platform around - but that's just my opinion.

The platform needs more native/browser-based applications and services that are easier to develop and maintain, more reliable, more responsive, better integrated, and look better than their Wintel counterparts.