(this is page 79 of 121)





Latest

On the data smelter

Any self-respecting ManGeek ought to love a term I picked up from the Economist a couple of weeks in an article on cloud computing: data smelter. Apparently this is moniker used for the huge data centers that Google, Amazon, Microsoft, Yahoo, and others are building on the banks of the Columbia river in Oregon. Located in the middle of the cheapest power available in the US, the name data smelter is a play on the aluminum smelters that peppered the banks of the Columbia over the last hundred years, but it's also great because it hints at one of the most relevant facets of the cloud computing/web services revolution: the ability for new services to recombine data hosted by other services in novel and interesting ways. We haven't even begun to feel the true power of how transformative this loose coupling of data and processing is likely to be; today's "mash-ups" are barely at the crawl phase of development in what we are likely to see.

And yet, it's worth pausing for a second to think about the cost of the current smelting. The Economist piece cites the Google data center at the Dalles as requiring the power of a town of 200,000 people. Most of this wattage goes to power the compute cycles that Google requires to index the world's information, and in most cases these cycles are well spent by running hairy algorithms that apply the bleeding edge of computer science to extract order from chaos.

But this is not always the case. For instance, at Tabblo, a meaningful amount of our general web traffic comes from Google Bot or one of its competitors. This despite the fact that we have well-structured RSS equivalents that could be polled/processed in a much more efficient way. Ditto for all of the much bigger user-generated content sites— they too have a meaningful amount of traffic coming from indexing bots while at the same time providing feeds that might provide just as much information for searchers while using less bandwidth, fewer CPU cycles, and not as much overall smelting.

The few times I've read any luminaries from Google talking about the semantic web in any shape or form (RDF, microformats, etc.), they always pooh-pooh it with slights like "people don't want to deal in angle brackets all day." And until I started thinking about the energy implications of these data smelters, I was likely to agree— after all, we're all still suffering from the CORBA/DCOM hangover of the last decade where a few vendors bamboozled the entire industry into thinking that an overwrought solution for remote process data exchange was the answer to all of these coupling needs (watch the WS-* offspring for a modern-day equivalent).

But last week Yahoo played a potentially game-changing move with its pledge to support the semantic web standards (microformats, RDF, etc.) across all of its properties. As much I tend to write off Yahoo as roadkill on the Google highway, it's clear that a few folks there are still doing good things for the net and the planet.

If the other industry heavyweights are goaded into following through, we may end up running slightly cleaner data smelters in the near future.

Comments Permalink

Though I am a huge fan of my recently acquired Macbook Air, having spent the last week digesting the implications of iPhone 2.0, and more importantly the SDK, I can't help but feel that the product form factor we have come to know as the laptop computer is about to get crunched by the smartphone in the same way that laptops have killed desktops over the last few years.

Let me start with the real strength of the Macbook Air: it's diminutive weight and slim shape. That these two attributes are enough to offset all of the other shortcomings (not nearly enough hard drive, barely enough processing, and no way to work for more than 3 hours at a stretch) says a lot about how much we care about portability these days. And yet there are still a whole bunch of things that are just too cumbersome and "disconnected" about the Air for me to feel as good as I should about a 3lb computing environment I could absolutely take anywhere. I want to love it but my heart has now been stolen by that lozenge-shaped computer traveling as a phone in my pocket.

"Stupid fanboy," I can hear you thinking, those crappy "smart" phones don't have the juice to run a real computing experience! For all who think that, I'd suggest taking a look at the iPhone SDK keynote, about 2/3 of the way through when the application vendors get paraded on stage to show their games/business apps/etc. built with the iPhone SDK. Ignore all of the Apple love and focus on the responsiveness of the UI, and the amount of "computing" taking place. Additionally, with rumors of the iPhone moving to Intel's new super-efficient x86 processor, I can only imagine that this will only get better— despite the need to protect that teeny little battery. And what Apple does here, you can bet Nokia, HTC, Ericson, and the rest of the 7 Dwarves will implement as well.

The other big criticism about the phone's ability to eclipse the laptop comes from the keyboard lovers— after all, no one could do anything other than consume media, send SMSes and reply to the occasional email on small keyboards, nevermind virtual ones like the one on the iPhone (which I am now ready to predict will win over the ant-sized keys of Blackberry like devices). Enter Nokia's Project NoBounds, an attempt to point to a world where our smartphones can easily pair themselves to larger displays and fuller input mechanisms (think how far we've come with Bluetooth headsets in the last 5 years and then extrapolate to displays). Were the gloried docking stations (anyone remember the old Apple Duodocks?) to become even as pervasive as iPod-equipped alarm clocks are today in most hotels, we'd have just enough infrastructure that most of us would be willing to take a chance and leave the laptop behind. And of course, you'll also come to expect one at work and a few scattered around your house.

Other trends that are going to push us in the direction of laptops being relegated to a niche similar to where desktops thrive today (CPU-intensive creative crafts):

Storage: For whatever reason, laptop drives have not kept up with the mushrooming demands of multimedia. I always buy the biggest drive I can get, and yet despite changing my laptop every 12-18 months, I'm always running out of space. So much so that most of my multimedia has already been split up according to file type— which is a horrendous PITA. I want a canonical copy of all of this stuff to live on some network-attached hard drive in my house with a backup in the cloud. Think Time Capsule + S3 with a nice usable layer of software to let me sync partial copies from my various portable devices. When this happens, we'll be able to really make the switch to flash for local storage (64GB won't seem small at all) which points again to the ability for the mobile device to become our primary computing device.

Closer physical contact with the rest of the world: No two ways about it, the ability to have a camera, location information, an accelerometer, and a net connection available within 5 seconds of taking a device out of your pocket constitutes a new platform whose real depths we have not really even begun to plumb. This is as true for frequent travelers as it is for people at the supermarket. And the funny thing is that even when these capabilities are brought on to the laptop to extend its term of service (witness the CMOS camera mounted on most portable computer bezels these days or attempts to add location aware functionality to laptops), they just don't yield as much use as you'd expect. Mobile computing is about user experiences that last 15-60 seconds and unfortunately laptops just aren't competitive under such stringent time pressures.

I love my laptops and have to really stretch my brain to think that in 3-5 years, I'll be looking back at this svelte Macbook Air in the same way that I saw the last desktop tower I ushered out of the house last fall. But it's coming and over the course of the coming weeks, I'm hoping to get some time to noodle on what the 15-60 second platform that will replace them is going to mean for those of us building the consumer apps of the next 3-5 years.

Comments Permalink

Fast on the heels of my energy-on-the-brain week in San Diego, I decided to run some experiments with my recently acquired Kill-a-Watt to debunk some myths about consumer electronics and power consumption. What follows is by no means exhaustive, but I figured I would write it up as it has frequently been the topic of lunchtime conversation at the office— with people arguing both sides of each argument as though it were politics and not simply electricity 101.

The basic statement that I was trying to confirm or disprove was that your computer/cellphone/ipod/etc. charger sucks electricity even when it is not connected to a device. Savvy environmental marketers have called this the "vampire effect" or the problem of "phantom power," and truth be told, after I first heard the term, I could never look at one of those cuddly black bricks the same.

KillawattSo I went around the house looking for as many bricks as possible, putting my Kill-a-Watt between them and the wall source of power and then connecting and disconnecting their associated devices. An aside: For those that don't know what a Kill-a-Watt is (pictured here), it's one of several cheap gizmos you can buy to plug between a given appliance and the wall to measure how much power is being consumed. I'm not quite sure how it works, but quickly testing it on both 60 and 100 Watt lightbulbs convinced me that it worked as billed.

The result: for each of the 13 bricks that I tried, ranging from a wireless phone charger to a MacBook Pro power adapter, the vampire/phantom thing is complete BS. The moment you disconnect the associated device the Watts measured on the Kill-a-Watt go right down to zero. Interestingly enough, this is equally true for low wattage chargers like the iPhone one (~1-2W while charging). It makes sense— after all I'm fairly certain that a fairly cheap circuit on the power adapter can get a good sense of load and just cut the whole power supply off if nothing is connected. As a funny aside, it seems that there is a whole category of "smart powerstrips" that are sold to protect the user against this bunk phantom power thing.

Since I had the Kill-a-Watt out anyway, I then went on to try to prove one of my own wacky theories: that you could actually conserve overall power when using your laptop at home by a) pulling the battery out when you were connected to the outlet and b) never leaving the machine connected after it was fully charged.

My theory about pulling the battery out went something like this: if a machine consumes X amount of power just to run, then also charging the battery at the same time must require X+Y power where Y is whatever it takes to charge the battery. And to point b above, even when the battery was fully charged, there would still be some amount of power required to trickle-charge, or keep the battery topped off.

However, in my very simplistic testing, it seems that with or without battery attached, my MacBook sucked in about 20-30 Watts pretty constantly (the variation being most directly affected by the display brightness). I tried it with 50% charge in the battery, 80%, and even 100% and it just seemed to hold constant in its power consumption in all cases.

Which of course leads me to wonder where the incremental power to charge the battery comes from. My current guess is that the Kill-a-watt does have some standard margin of error (that I didn't see with my lame lightbulb tests) and that this is where the incremental power to charge is being missed.

To end on at least one energy saving tip, for a typical MacBook in sleep mode, the cost of trickle charging is about 1-3 Watts which means that unless you are going on a long airplane ride the next day (or are a power hygiene freak who must have all devices at 100% charge all the time), you will save a little bit of power by pulling the adapter from the machine (though not unplugging it from the wall) when the little charge light turns green.

Whew, ok. With that behind us, I promise to get back to InterWebs topics tomorrow!

PS, if anyone knows of a relatively cheap device that can act as a Kill-a-Watt does but store data points over time for subsequent analysis on a computer, please let me know. My next set of experiments requires more than just eyeballing the display as the junk attached sucks power.
Comments Permalink

What I learned at Etech

As I sit here writing this sentence, my brain is consuming about 20 watts of energy. For some reason that seems like a lot to me, until I think of the fact that 3 of me could run on the energy that one of the many 60 watt lightbulb I constantly forget to turn off.

Thought exercises like this one were the natural outcome of attending Saul Griffith's energy literacy keynote— by far and away the best of the keynotes I attended. Underpinning the talk was a relatively simple premise: to put all of the things we do on a constant energy scale (watts consumed) and check out where we might be terribly wasteful (in my case: airplane travel by an order of magnitude, followed by commuting to the new HP offices, with buying too many gadgets as a distant third). His talk was fascinating because of this one act of normalizing data— and has given me a much better view of things than Al Gore's documentary or the various overly complex papers and books I've tried to digest. Watts, watts everywhere (edit: for clarification purposes I should say that a watt is a unit of flow, the real absolute unit for energy is the joule; 1 watt is just 1 joule per second [thanks Paul in the comments)— for everything from running your gas sucking automobile to that can of Coke you are about to drink.

And speaking of wattage, that brain of mine seems pretty efficient compared to the laptop I'm writing this on which sucks up about 40 watts to give me an infinitesimal amount of the processing my puny brain puts out (which according to Kathy Sierra's self-help /feel-good keynote could be vastly improved by less Brain Age and more intense periods of concentration). Why then does it suck up twice the power? According to Stan Williams, from our very own HP labs, because we may be using an inefficient model for computation. In a fascinating talk that made me proud of HP labs, he gave a survey of techniques used to increase computing power to get to computers whose power nomenclature sounds like the Sunday morning special at Waffle House (ExaFlop?). I've never been a big fan of the quantum computing stuff (which according to Williams only has application in crypto, if at all), but I was riveted by his assertion that our starting assumption that computers ought to be built with standard boolean logic (AND, OR, NAND, XOR) might be something that we want to check, especially because we trace its origins back to some 16 year old farm boy named Claude Shannon (famous for his work in information theory) who was getting his PhD at MIT in the 30s. The best part of Stan's talk: how even keeled and practical he sounded about all of the crazy sci-fi shit that is getting done in his lab! This is one guy I am definitely looking up when next at the mothership in Palo Alto.

Flawed boolean logic notwithstanding, it was awesome to see all of the hardware hacking taking place at the conference. Bug Labs and Chumby are neat attempts at building companies out of some of this creative energy, though limited precisely because of commercial constraints. Much more interesting were a number of conversations I had with folks about Arduinos and the whole related ecosystem of camera/GPS/network modules that are springing up around them. The king of cool in this category was definitely Chris Anderson (of Wired fame) with his DIY Drones talk on making full autonomous flying things out of cheap componentry. If I hadn't seen his videos, I would not believe that the stuff in his talk was even remotely achievable by an amateur. Particularly cool was the UAV that he flew over the Google campus to take pictures that he was then able to collage through GPS coordinates captured during flight.

And speaking of GPS, geo was the new black at this year's conference. The number of talks, hallway conversations, and keynote mentions that centered around location hacking were too numerous to list here. It is honestly beginning to feel like understanding how to work with geo-data is the new SQL with this crowd (where is the O'Reilly book on this topic?!?). And amidst this buzz, Yahoo launched FireEagle, a thin piece of geo middleware that lets users register producers and consumers of geodata. Despite the panning the poor guys took on TechCrunch, what they are trying to do strikes me as super cool and important (email me if you want an invite).

It's funny— during the first 12 hours of the conference (which I have to admit I was forced to duck in and out of due to work obligations), I felt ready to declare that it had jumped the shark (especially after Tim's rambling and slightly depressing opening talk). But in the end I'm glad I went, mostly because of the way that an event like this helps to remind me of the for some folks, it is a core belief that anything is hackable (including according to Lesig, government).

More than anything else the crowd that O'Reilly is still able to attract to this event exudes that ethos from every pore. Magazines like Make, blogs, and websites are helping to propagate this more broadly, but seeing it first hand— even if once a year— helps to recharge the batteries that run this 20 watt brain.

Comments Permalink

En route to ETech

Heading to Etech this week which should be a great treat not so much because of the high octane crowd that O'Reilly attracts to this conference but because it seems like with Etech 2008, the conference organizers are taking a turn away from the standard Web 2.0 themes and looking at other places where technology is making a big difference in how people relate to the world (go read Tim's perspective on it).

I am excited about the green-tech theme that seems to be all the rage for the software/Internet people these days, and as a result, I'm looking forward to all of the sessions along that track, starting with the energy literacy keynote. In fact this morning I was reading a piece in the Economist on how all of the prominent geeks in Silicon Valley are becoming green tech entrepreneurs which made me wonder whether in technology there is such a thing as the entrepreneur truly divorced from the type of innovation wave which he exists in. During the explosive early growth stages of the Internet there were a whole bunch of entrepreneurs who came out of PC-era companies and did ok in their 2.0s, but it bears mentioning that there was none of those second timers really knocked it out of the park, and that today's Internet powerhouses are all led by folks who had none of that historical baggage/experience with them.

Nonetheless, it is the rare conference these days that offers the chance for a broad and fresh perspective— which is exactly what I'm hoping Etech will be this year.

Comments Permalink