The IT Industry's Red Shift Theory

Stony Stevenson writes "Sun Microsystems' CTO, Greg Papadopoulos has come out with a Red Shift Theory for IT which posits that an 'elite group of companies are consuming inordinate amounts of IT infrastructure, well beyond most other businesses, and that their demand is growing exponentially. This trend, Papadopoulos maintains, has implications not just for IT's most insatiable consumers, but for the structure of the computing industry itself. It's not just about how many CPU cycles a company uses. Papadopoulos argues that red-shift companies will enjoy exponential business growth in the coming years. Blue-shift companies — those whose processing needs aren't exploding — will grow at about the same rate as GDP, he says.'"

so.......

A hardware company says that buying more hardware is a good thing for your company. News at eleven.

(Yawn).

But... Doppler...

But they used the Doppler effect [wikipedia.org] to explain it, surely that little scientific reference counts for something?

Re:But... Doppler...

Maybe if they'd used it correctly. But their use of "blue shift" is completely inconsistent with the metaphor.

(A blue shift happens when things are getting closer to one another, not further apart. These supposed companies not accelerating would look exactly as red-shifted to the accelerating ones as vice versa.)

Re:But... Doppler...

I believe he is comparing them to the GDP.

Re:But... Doppler...

Yeah, it should be a black shift - where your company stops buying lots of expensive Sun hardware and shifts from being "in the red" to "back in black". ;).

'Exponential' fails common sense.

Exponential growth for a business isn't even possible, unless maybe you start out really, really small, and for a short amount of time.

Exponential business growth is not possible simply because the number of customers is finite. Even if you make something that all 6 billion people will buy, you're still not going to be able to maintain exponential growth very long before everybody is a customer.

Re:'Exponential' fails common sense.

Even if you make something that all 6 billion people will buy, you're still not going to be able to maintain exponential growth very long before everybody is a customer.

Not necessarily. You can grow the number of customers by a small amount, but increase the number of things you sell them. Take a look at Apple. While previously they got most of their money from Macs, now most of it comes from iPods and iTunes. If they sold everyone who owned a Mac an iPod, then they doubled their sales without increasing their customer base at all. If they can then sell them an iPhone as well, then they have grown their business even more without increasing their customer base.

Of course, eventually you need to be producing everything people use and consuming all available resources. Then, if you want to keep growing, you need to start doing something really novel (asteroid mining, that kind of thing), but that won't happen for a long time. Exponential growth is completely possible for short periods; Sun experienced it themselves in the .com boom.

Re:'Exponential' fails common sense.

Exponential growth for a business isn't even possible

Sure it is. 2% growth is exponential.

Re:no, no, no

Come on, you're oversimplifying. He also says slower growing companies grow more slowly. It's that little bit of extra insight that makes the difference between ending up as CTO instead of a janitor in the modern corporate world.

So ...

Do the "red-shift" companies employ a "blue ocean strategy"? :-)

If Professor Hawking says it...

it must be true. Why else would they allude to a phenomenon of cosmic proportion to describe a simple "growth equals growth" observation?

Huh?

One shift two shift, red shift blue shift?

Red Giants

The problem then is that red giants end up as white dwarfs, and no-one will do business with a dwarf!

So companies that innovate will make more money?

Get out of town...

It's true, innovation takes a fair amount of software development (if that's what you're building). If you're doing accounting systems for state governments, your revenue will grow more slowly, unless you add a new state or something.

Of course, out in the real world, there's more than IT - corning has been growing fairly fast over the past 5 years, and they seem to have done it with carbon filters.

Every week

Every week, yet another IT business-man / manager /columnist which fondly remembers the Internet Bubble comes up with yet another theory on how IT/Internet/Networking is causing/will cause an infinite boom of growth and properity and those which do not jump onto the train of [fill in something he's trying to sell] will be left behind in the dust.

I thought the Bursting of the Bubble had cured people of falling for this kind of arguments (which were used over and over again during the Bubble to justify insane valuations for companies which never made a cent).

Guess the leeches didn't gave up on trying to suck the fools dry yet.

Re:Every week

As somebody who does software for a living, i can tell you that the biggest factor in the success of a company is ... business processes.

In the big picture of how a company is successfull, software is a tool, networking technologies are tools and the Internet is a tool. It all boils down to people and organization - individuals and the way they work together.

IT is something that can fit into the business and can empower the people to work beter - it's not a silver bulet which will magically transform a mismanaged company into a growing, thriving business.

Sorry to burst your bubble, but from someone that has been working in IT for many years, across several industries, it's my experience that the best success stories are not "IT transforming companies", instead they came from "companies that mold IT to their needs".

Every company I've seen is still mired in red tape and completely backwards use of much of the technology we require to survive.

Red tape is an organizational problem, not a technological problem. Bringing IT in without solving the underlying problem will just result in adding new layers to the bureaucracy (been there, seen it happen, not a pretty sight).

The truth is, IT brings with it whole new time and money sinks (license control, networking/systems administration, IT security, software development and costumization, outside consultants, etc) which would not be there without IT. In truth, as many of us in IT have seen again and again, often the blind, vendor-pushed, fashion-following approach to deploying IT in a company results in wasted money and a decrease in productivity (for that company, the vendors are probably quite happy).

During the Boom years, many companies where managed by people that did not understood that technology is a tool for empowering the business, not the other way around. Countless managers let themselfs be taken with sentences such as "utilizing technology to its fullest potential", "software that can boost our capabilities even farther", "the technology we require to survive", "streamline our business the way software is capable of doing" and other such sales pitches and so let their companies be taken for wild rides, where the only ones that really profited where the vendors. Hopefully, most learned their lesson, and the latest generation of of CxOs is beter at separating the technological wheat from the chaff.

PS: Even though i'm someone which if often brought in to help clean up the mess done during one of those "wild rides" (said mess having been done often enough by the unholy association of software vendors and IT consultancies), I would much rather loose that part of my work in the future that be faced again and again with the kind of raw sewage which is all that has been left from the blood sucking feeding frenzy done by the above mention vendors and consultancies.

PPS: Yeah, i'm sour about this.

Cause and Effect

This is backwards. Companies which are buying hardware are buying hardware because they already have a successful business model (or one they expect to be successful). The differentiator between successful and unsuccessful companies isn't how much hardware they buy, it's the viability of their business plan/product.

The ST Industry's Red Shirt Theory

'Sun Microsystems' CTO, Greg Papadopoulos has come out with a Red Shirt Theory for IT which posits that an 'elite group of aliens are consuming inordinate amounts of ST infrastructure, well beyond most other aliens, and that their murders are growing exponentially. This trend, Papadopoulos maintains, has implications not just for ST's most insatiable consumers, but for the structure of the space industry itself. It's not just about how many warp cycles a company uses. Papadopoulos argues that red-shirt companies will enjoy exponential deaths in the coming years. Blue-shirt companies -- those whose heads aren't exploding -- will live at about the same rate as GDP, he says.

Bullshit

Basically, spending exponentially more each year on hardware will give you an advantage. True if by "advantage" you mean "become bankrupt before your competitors".

TWW

"Sir, we need to upgrade our infastructure..."

"Sir, we need to upgrade our infastructure..."

"Why is that, Jenkins?"

"According to a new report, our profitability is directly related to our processing needs, sir."

"Well, I'm not seeing how our needs have increased enough to go over our yearly budget for hardware, Jenkins."

"Sir, having extra processing power will increase our productivity..."

"How's that, Jenkins?"

"Porn."

"Make it so, Number One."

Inventing Terminology for CEO's

Don't be harsh. Proactively leveraging six sigma synergies in a blue-red shift discontinuously changing ecosystem leads to a re-efficiently contracting contingency workforce paradigm tending toward the rightsize.

Only a fool would ignore the gainshare effect of empowerment strategies that insourced intellectual capital reallocation due to lean Kaizen open door management obviously creates. The mosaic effects of capital market global forward-trends account for fully 4 points share of Sun's complex-component earnings per diluted ownerstake last quarter.

With two new colors to work with, things can only improve.

Re:Inventing Terminology for CEO's

There's a neat trick you can play with Markov chains to generate this sort of text. It may not be as good as your handcrafted version, but it makes it easier to generate larger texts.

The algorithm basically works by feeding in a sample text, from which you generate a statistical model of what words are likely to follow any given N-word sequence. Then you select at random which of the possible suffix words to output given the previous N output words. (Obviously you need to provide the initial conditions, i.e. the first N words of output, to the algorithm). If you allow punctuation to be considered part of a word, it seems to produce reasonably grammatical sentences too.

Picking N=2 seems to work pretty well. I imagine if we fed in a bunch of buzzword-laded management texts we'd get some great results.

Re:Inventing Terminology for CEO's

You know, it was annoying when marketing goons made up jargon. But it's more annoying now that they steal technical words from other fields and misapply them.

Pseudoscience

I no expert on this One Shift Two Shift Red Shift Blue Shift business, but where exactly is the author standing that the GDP is approaching him, and how many terahertz is the IT industry is oscillating at?

Every company is racing away from every other

And the rate of expansion seems to be increasing from the little-understood Dark Capitalism.

Microsoft has BSOD

...so Sun will invent RSOD.

Its really simple to understand....

The more digital consumer babeling and viewing gets the more processsing power you need to process it.

A lot of the shift is from a transmission media that didn't need a lot of processing power to one that does.
Where the amount of media consumers access is growing at a rate less than that.

Where is the difference in the eyes of the consumer... their wallet mostly.

But we have more choices now and a greater level of convenience. IS it worth it?

There using SlashDot to test Marketing

They are using slashdot to test marketing. Now how they score articles on wether it gets more or less vitriol from us slashdotters is up for debate.
Remember, Marketing, or advertising targets only the week-minded, sane strong willed people will do it themselves.

All I got from this:

1) Some people buy more fast computers than other people.
2) Like, a *lot* more.
3) These have to be bloody *fast* computers, if they're causing Doppler shifts.

Let's for a moment ignore how poor this analogy is

Ok actually let's not. Let's play a little imagination game.

Picture the whole of IT companies and companies that use IT as a spherical microcosm. The normal physical dimensions in this microcosm are instead dimensions of economy, scale, and technology. Like our universe is puportedly doing, this sphere is expanding. Now, because he claims these "elite" companies are basically exponentially distancing themselves from the other companies, they could arguably be on the extents of the expanding sphere. Although there are fewer of these companies they require much more of our microcosmic IT space. Fair enough. Moving away from us, and red shifted from our perspective in the exact center of the sphere.

So far, so good I guess. Then we have blue-shifted companies. In this microcosm, in order to become blue-shifted to my perspective, you'd have to be heading away from using more advanced technology and also using less of it. Frankly, I can't come up with a single company in the world that's doing that and is successful. If you don't look at in on a time-scale of a year or two, but rather of fifty, one hundred or even five hundred years, it's pretty damn clear that every single company is red-shifted from the perspective of an individual in the center of the microcosmic sphere.

End game.

So I put it to you, /. reader, to fix my broken visualization of his analogy. If you can't, I have to say to Popdopopdoloslous, you have lost me.

TLF

1960's claim repackaged - why it faulters

The notion of utility computing, of course, has been around since the IBM mainframe. What's new is not just an explosion in demand from companies with mountains of data to climb

Not really. Data chomping has always been a desire. It just used to not be cost-effective. People indeed did think that IBM would sell bigger and bigger chunks of plug-and-play processing. However, minicomputers and microcomputers ruined that goal. It appears to go in cycles where centralization ebbs and flows as new technologies and tech fads come and go, changing the dynamic.

...Referring again to the power grid, Papadopoulos uses the analogy of a massive, very hot power plant that produces multiple megawatts of energy versus an array of portable Honda generators....But the jury is still out on whether the big iron from Sun will win out over arrays of dozens or hundreds of commodity boxes. "We'll see [an exponential] increase in the number of servers sold, yes," says technology pundit Mark Anderson, author of the Strategic News Service newsletter. "Will they be Sun servers, running Solaris? I'm not so sure. Everything is headed toward open systems, mostly Linux on commodity servers."

This is exactly what the mainframe-centric model proposed. In a way, it is what economic communism claimed: that centralization factors out wasteful duplication, and allows economies of scale. However, reality has shown that to be messier and more difficult than it looks on paper.

One of the counter factors is that it is often easier to do massive data chomping on a desktop PC. For example, Oracle servers are almost always overburdened and therefore slow, so that people tend to download raw data to MS-Access and similar tools and process it themselves faster. If the Oracle DBA's pumped up the power of the Oracle servers, then people would use up all the extra power pretty quick. it's like building freeways: the more you build the more houses are built. Traffic bottlenecks are what limits many cities from growing further.

Similarly, if one manages their own processing (via MS-Access-like tools), then they manage their resources more efficiently. The Oracle DBA cannot really be the one who decides who is wasting resources or not because they cannot keep their head in everybody's business needs to see if they really need an 8-way join for an odd marketing report, etc. Localization thus polices resources better since one is competing against themselves (their own desktop or local department/office servers).

Crunching power is no value in itself

Contrary to what this theory says, computing power has no value in itself. I can think of no scenario where one company wins over by simply 'outcomputing' competition. It may be that global companies with a versatile infrastructure and the neccesary computing power to control it may have a competitive advantage - but that's just like saying Coca Cola has a competitive advantage over Bobs Village Brewery. Killer App Software are built by small teams of knowlegeable people on affordable hardware with relatively standard processing power. And however well that scales is mostly dependant on having the right people with the right knowledge in key IT decision-making positions (rather than having a higher cap on the IT budget). A thing more unlikely to happen in larger companies.

I actually would think it's the other way, with large operations wasting huge amounts of CPU cycles because of bizar IT-business decisions and thus requiring more of it per dollar of turnover than smaller companies. A prime example that play in the same direction for this is the developement of the telecom market in Germany in the last few years. Within just a few years all companies have gone from wide ranges of Über-complex billing rates and tarif models to simple phone flatrates. This trend was started by new kids on the block like Arcor who quickly figured that the billing infrastructure would be far to expensive to compensate for the potential gain in income it would provide. So they introduced phone flatrates which are by orders of magnatude simpler to bill and process. And others followed suit.

Bottom line: The theory sounds interesting, but I don't think it sticks. If it does, then only under special isolated circumstances.

A More Applicable Metaphor

Since those companies that are growing more slowly are still growing the same direction as the "red shift" companies, there is no blue shift, just different shades of red. This is not a very good analogy.

A much better one would equate companies with countries, and growth with energy useage. There's one company that's the US, growing/using much more per capita (per dollar) than the others. There's a few that are growing/using somewhat less, but in the same order of magnitude. There are more and more as you go to the low end of the list, with many not growing much more than the GDP (equivalent to using only their own energy production).

Of course no CEO wants his company equated to something that squanders resources and dumps crap into the environment in the process. No sense of humor.

Growth and Acceleration as a metric is flawed

I cannot seem to understand why people cannot see this flaw as easily as I do. Success of a company is often measured by its growth rate and indeed by its rate of increase on the growth rate (acceleration). There is a limit for EVERY market. There is a saturation point for every market. And when the health of a company is measured not by its stability or state in the market place but by "growth and acceleration" I have to wonder what drives the mentality that it's actually a good idea outside of what it does for those who buy and sell stocks on the market. (So yes, that's exactly what it's all about... duh)

So when did we all lose sight of what is good for a company? Matters like quality and customer satisfaction are no longer a consideration? And every time I hear "this company is buying that company" or "we're on a growth surge" or some other such nonsense, I have to wonder why anyone would think this sort of institutional business instability is a good idea for anyone except those who play the stock market?

Will we have to suffer another great depression before people realize that the cause of so many of our business, labor and national monetary health problems are rooted deeply in the short-sighted notion that whatever a business does it should be as a means to provide value for shareholders? I think the answer is yes because short of a disaster, people will have little motivation to see where this all leads and turn around before it's too late. And unfortunately, while one person might catch a glimpse of the future and become more sane, the people who are still insane will consume him as a means of satisfying their growth strategy. I get the mental image of a bunch of cannibals strategizing their own growth and acceleration success plans in how to consume their environment. The logic is rather unsettling to me.

Repeat after me...

There is no correlation between IT spending and productivity or profitability.

This is the same old saw that hardware firms used in the 1980s-1990s. Gartner used to say you should spend on IT as a proportion of your revenue.

But numerous studies [infoeconomics.com], based on publicly available data, debunk that view as bullshit.

It's not that IT is bad -- it's just that you have to blame or praise management for the proper application of it. Which is just another way of saying "you can't spend your way through problems without thought".

NOW, there's a valid argument here, but it's a lot more subtle than the bylines. One has to dig into Papadopoulos' quotes to get the jist of this as: "you should have the management insight to take advantage of the inherent cost savings that are due to Moore's law." This has been hampered for decades due to inflexibility with the software -- something that virtualization and utility computing is seeking to fix, and an area that Sun wants to compete in. Indeed, this is a big deal.

But it doesn't mean your computing needs will skyrocket, unless your management has an insightful, productive application for all of that power. Google does (selling advertisements along side day-to-day networked computing needs), but I'm not sure the rest of the Fortune 500 has turned to apply that level of creativity to their situation.

Hold on, something about this doesn't compute

IT assets cost real money. If a company purchases large amounts of "CPU cycles" it requires a large expenditure. How that translates to a faster growing business is unclear. Maybe if we rephrase the source article a bit it'll make more sense?

Let's try: "Company X says that buying lots of company X products will cause your business to grow at an exponential rate".

Aha! Now I see - it's just more marketing doublespeak. I wonder if you could get Sun to put these claims in writing?

metaphor alert

What kind of stupid metaphor is that? Red shift means that they are moving away from the observer fast, blue shift means that they are moving towards the observer fast.

As for the "theory", mainly what you're seeing is that fast growing companies buy a lot more hardware; they do that for two reasons: (1) they are adding customers and (2) they don't have time to tune their software, so they are using disproportionate amounts of resources. But the cause is fast growth, and buying lots of hardware is the effect.

Oh, one thing that's pretty clear: many of those companies are not buying Sun products. Why would they?

hype or space?

There are, at minimum, some companies which need massive amounts of computation, which needs to scale not just with their markets, but with their product development - where their edge largely consists in finding useful things to do with massive amounts of computation before their competition scoops them on whatever the next idea to take off will be.

Yes, in an arms race you often end up with in the silly spot of "But what do we need 10,000 nukes for." Well, what you need 'em for, if you want to be the only superpower on your globe, is to hopefully make the competition quit before they begin to seriously challenge you. Eh, bad analogy. The point is, in this "arms race" the "nukes" also contribute essentially to your products. And the people building your nukes had better be fully aware of the degree of your need, and competitive with the other armorers.

Sun, first in fusion.

Isn't the GDP exponential?

I would have thought the GDP was exponential since money tends to breed more money.

Even if it wasn't, wouldn't the presence of exponentially growing companies force the GDP to go exponential after a while. Maybe such companies will quickly die before they get big enough to do that, but a bad business strategy that leads to a quick death sounds like a bad business strategy (modulo the Enrons and SCOs of the world).

How nice! IT resource envy!

Sounds like a lot of things coming out of the Democratic Party and their underlings....like the USA is using "too much" of the sunlight? That was just a week or two ago.

Sheesh...well, "Divide and conquer".

And the there's the...

brown shift theory, where any and all overblown theories about the interwebs invariably approach utter bullshit at an increasing rate.

GDP grows exponentially

http://en.wikipedia.org/wiki/List_of_countries_by_ GDP_(real)_growth_rate [wikipedia.org]
all the numbers listed in that wikipedia page are the k in the exponential differential equation dx/dy = ky.

I sense a new Dilbert comic coming soon ...

... with Mr. Papadopoulos as the Pointy-Haired Boss[tm].

Umm...

Oil companies grew exponentially for awhile...then levelled off. Then cars were invented and they grew exponentially again...then levelled off.

Car companies grew exponentially...then levelled off.

Every new thing grows like this, then levels off to more or less population growth. (With new smaller exponential growths as new markets, e.g. China, open up...)

Computers, software, etc. are just more of the same. But we've only been at it, large-scale, for 25 years. 25 years after Ford started, you could still get into cars pretty easily. 25 years after oil, mass-produced cars hadn't even been invented yet.

$$$ != Management

I was recently approached by a neighbor who run a small mid-term rental company, meaning that they rent things for 1-30 weeks). His business is growing and he is having a hard time tracking and billing for all his assets. He thinks a little computer program could really bail him out of his grwoing problems. According to sun, his business will be saved if he buys some hardware.

I don't think hardware or software will save his business. He needs to hire someone who can manage his assets. They could write every asset on its own 3x5 card and record a rental on each line of the card. Every week, they just go through each 3x5 and generate their billings.

I personally would use software, but if you can't think of how to do it with pencil and paper, you have a problem that is more fundamental to your business than the software can solve.

Re:Oddballs.

And vat ever happened to plain old lavender blue dilly dilly? Silly.

- Prof. Von Drake.