I was talking to a good friend recently and reported some hearsay about how a server now costs more in its useful life than it costs to buy. I found that amazing, but his response was even more astounding. “Well, we should put them in poor people’s houses to give them heat,” he quipped.

It sounds dumb at first, but really, it’s pure genius. If that much energy is being used, and half of that energy is used for cooling, we could put those servers to work as electric heaters. The “host families” could also get some broadband access, and institutions would save on data center build-outs. It’s a shame that our culture and the technical practicalities of distributed computing make the idea impractical.

But it got me thinking. How much energy really is burned in those big data centers? What follows next is guesstimation and inference based on popular opinion and, er, Google search returns ( I may appear to pick on Google, but it’s just because it happens to be a convenient example…)

• Google is rumored to have anywhere between half a million and 1 million machines in data centers. around the world. I am assuming it is the largest single-purpose commercial installation that we know about. (Lets not think about the government’s data demands for now.)
• Each machine consumes about 500 watts of energy, including cooling systems.
• Energy overhead for networking and other support structures is nominal, so I’ll ignore it for my guesstimate.

So, let’s take the worst case here:

1,000,000 Machines using 500 watts of energy an hour = half a gigawatt an hour.

Wow. That’s a lot. In Google’s own words, that’s about half what a city the size of San Francisco needs every hour.

That poses a worrying thought: Information on the web is increasing in an exponential manner, and Google will increase its capacity to meet demand. Even doubling that energy use would require the kind of power produced by a mid-size nuclear reactor. Now, I’m sure my calculations are probably a little over-dramatic, but really the quantities are on order and kind of astonishing when you think about it. And Google is not the only one. The cloud computing craze has to be powered somehow and the cloud’s power will come from a huge collection of these datacenters.

We could “Google” less and index only some of the world’s information — but heading back to ignorance doesn’t seem to be the right path to me. Instead, what we need is to rethink the “faster, better by throwing in more power” mentality of processor design and think around the physics of current computing and energy supply. It’s no easy feat. Perhaps also we need algorithmic innovation as well (Green PageRank, anyone?) Google is stepping up the mark with its investment programs in sustainable clean energy and building closer to energy supplies.

Collectively, chip designers, programmers, users, policy makers and academics will have to create a gestalt of contributions that run leaner, cleaner and cooler. Perhaps they should look at Google’s lead as a start

Note: Katie Fehrenbacher, editor of Earth2Tech, will be leading a panel looking at ecologically better platform solutions at our conference Structure 08.

Thanks to Josh Aller, James McBride, Saul Griffith and Aaron Huslage for their assistance.

Related research and analysis from GigaOM Pro:
Subscriber content. Sign up for a free trial.

• Why iPad 2 Will Lead Consumers Into the Post-PC Era
• The Near-Term Evolution of Social Commerce
• Content Farms: The Players, The Benefits, The Risks

We speak faster than we can type. Much faster. The average conversation speed of an American speaking English is 120-160 words per minute, whereas the average typing speed requirement is about 40 words per minute and the top 10 percent of keyboarders type at speeds of 64 words per minute or faster. Do the math. You could communicate nearly three times faster with a computer if it could understand you.

So why on earth are we still using two-dimensional representations of the real world like desktops, menus and pointers? I believe we could do a lot better by exploring how applications can work better with voice-driven user interfaces. Not just pure voice interfaces but also what are known as multimodal interfaces, which combine the best of vocal/audio interface elements with the best of what the screen can do.

Voice-driven user interfaces have long been the holy grail of human computer interface design. Science fiction is a great representation of this – in our fantasy worlds, characters interact with machines by speaking to them, not by accessing some desktop start menu. (Remember Scotty in Star Trek IV: The Voyage Home? “A keyboard? How quaint!”)

Unsurprisingly, Microsoft (MSFT) has been slowly but surely forging ahead in this area for a long time now. It started with massive investments in the ill-fated Lernout and Hauspie, whose founders pioneered a lot of speech interface technology. Microsoft built the technology into various facets of the Windows operating system, and although most of the functions were never widely adopted by the general public, they still remain there today.

Its recent acquisition of Tellme Networks was eye-opening. Most people saw the acquisition as a way for Microsoft to start bringing the other killer app of business technology — the telephone – into a bigger offering where Microsoft Office was integrated with voice-based unified communications.

I think, however, that Microsoft is smarter than that. When it comes to speech, speech interfaces and actually using speech interfaces to make money, Tellme has one of the best teams in the world. They helped push and standardize the VXML format, which helped make telephony developer-friendly, and they released an awesome Mutimodal local search app as well. More than any other company, Tellme will know how to scale from the intranet and Internet to the worldwide telephone network. Don’t forget, there are about 400 million PCs in the world and 2.2 billion telephone handsets. That math wasn’t hard to comprehend in Redmond.

So what, potentially, is next? What are some of the ways that voice-driven user interfaces could change our computer experience for the better?

Idea 1: Help mechanisms. Rather than presenting me with thousands of options in a help menu — none of which have anything to do with what I need help in — just connect me, voice-to-voice, with an expert.

Idea 2: Related conversations. Let me start a conversation around a document while I’m reading it. Or give the document a persistent chat associated with it. Failing that, the back end could organize a meeting for those involved in the doc and make that happen. After all you want to discuss that doc, right?

Idea 3: Smarter transcription. Transcription, in other words, that takes into account a person’s cultural background and doesn’t solely interpret sound as though it was coming from someone whose first language is English.

Idea 4: Coordination commands. “Open the memo Steve wrote last month, add the comments from the last document we worked on and e-mail them out to the people in this spreadsheet. Make sure there is read receipt for everyone and schedule a conference call between them for the 24th.” What do you suppose the sequence of menu commands and clicks would look like for that?

We can improve other activities through voice technologies as well, such as browsing, searching and general UI. A conversational (discourse-driven) approach to
searching for documents would be particularly useful and rewarding. I don’t know what Google (GOOG) is up to in this area but with its unholy brain trust the company must be looking at this, especially if it’s starting to build out its rumored GPhone play. And take a look at the console games industry, which has used voice to make games more fun.

We also need a lot of other things, like common sense technologies and artificial intelligence, to make really good user experiences, but I’d encourage software manufacturers to start bringing in simple but useful speech elements to their user interfaces so we can start to bring them to the UI vocabulary of people out there. Just as long as they’re brought in with usefulness in mind, not as an afterthought.

And in an effort to practice what I preach, this article was dictated using DragonDictate (now known as Dragon NaturallySpeaking), which I can assure you is damned accurate, with only a few changes required. Without the acceleration that it provides me — especially with my two-fingered keyboard skills — neither this article nor the obscenely full e-mail inbox I have would ever get any attention. I only wish it could proofread for me.

Ah well, maybe one day.

Related research and analysis from GigaOM Pro:
Subscriber content. Sign up for a free trial.

• Why iPad 2 Will Lead Consumers Into the Post-PC Era
• The Near-Term Evolution of Social Commerce
• Content Farms: The Players, The Benefits, The Risks

Over the last few months, it’s been Facebook’s turn. The site’s rise in traffic is in no small part due to developers being enabled to deliver their apps using the Facebook API. Some of these apps are getting serious traffic; see the Alexa results for the ex-PayPal guys over at Slide.com, for example. But some words of warning about social applications.

By Surj Patel

As Friendster found out the hard way, exponential network growth is compelling but the hangover can be deadly. In simple terms, your audience can grow linearly but your computational requirements often grow exponentially (social network computation has a lot of what’s known as NP-complete problems). If your app on the Facebook platform succeeds, you don’t then want it to die on the vine for lack of server resources. You need to make it scale and you need to make that happen quickly. And of course, cheaply.

While the guys in Palo Alto are happy that you use Facebook as your platform, they are not going to be happy if, as you succeed, you take them down with you. The techies at Facebook have been careful to ensure that they offload as much processing to you as they can — which means that in order for your app to scale you need to look at buying plenty of machines which, in turn, means lots of money.

The problem with traditional hosting resources is that you need you to pay up front, then they either hammer you with overages for being successful or charge you up the wazoo for unused capacity that just generates heat. What you need is what the world of suits would call “just-in-time resources,” which can be roughly translated into: “I get what I need, when I want it, and I pay only for what I use and no more.”

But of course you’re broke, sleeping under your desk and eating ramen noodles. So what can a poor entrepreneur do?

Welcome to Amazon and S3 and EC2 — processing power (EC2) and storage (S3) on demand. These services let you access computational power and storage only when you need it and, better yet, pay only for what you use. The last time I checked, it was 10 cents an hour for the server, 10 cents for every gigabyte of data written and 18 cents per gigabyte read out – all for a virtual box with 1.7Ghz x86 processor/1.75Gbytes of RAM/250Mbs of bandwidth. Nor are you limited to one usage; use as many as you need or want and can afford. But get ‘em while you can, as a waiting list has appeared in the last few months.

So where could you use such “meta services” when developing an app for Facebook? I talked to Jeff Jolma, the expert Facebook programmer at Snapvine, which has been adding voice services to Facebook and MySpace for the past couple of years. He said they’re great for getting the job done and making sure the job keeps getting done. “You have to be aware of some of the limitations with meta services, but if your app can work within their parameters then it’s a cheap way to scale up and down as needed,” he said. They’re also “definitely worth looking at” for Facebook apps, he said. Other uses for EC2 and S3 include:

* Images – if you have to store or process images beyond Facebooks provision
* Video storage and processing
* Heavy storage – DB replication, video, backups, transfers
* Backup – safe and secure
* Batch processing – large amounts of cycle-intensive processing all at once
* Data mining
* Spike demand handling (the Slashdot antacid)

And many, many more.

Now you may read this and think, “While it’s a great idea, the cost to develop the software to handle this would negate any benefit, right?”

Wrong.

Someone has probably already probably done most of work for you; use Google to find them. There are various meta services out there that can help you as well. Some, such as with WeCeo and RightScale, are selling computation services through a web site/API and will manage everything for you.

If you prefer to roll your own, go directly to the source with EC2 or S3 (for related ideas and support go here for EC2 and S3, respectively).

So go on, give it a try — you really have nothing to lose. Make yourself and the young Mr. Zuckerberg richer. While you’re at it, try to have some fun and please do share what you discover about meta services, be it good or bad. We’d also love to hear about services other than Amazon’s, so please post your findings below.

Thanks to Jeff Jolma at Snapvine for help in researching this article.

Related research and analysis from GigaOM Pro:
Subscriber content. Sign up for a free trial.

• Why iPad 2 Will Lead Consumers Into the Post-PC Era
• The Near-Term Evolution of Social Commerce
• Content Farms: The Players, The Benefits, The Risks