Energy Savings, Strange Attractors, ...

... the Intrinsic Cost of State Change, Orbiting Alien Voyeurs, and 200 Square Kilometers of Solar Panels Somewhere in Texas

The Silicon Valley Leadership Group and Berkeley National Labs recently published the results of their first Data Center Demonstration Project (pdf). (Disclosure: My colleague Teresa Tung of Accenture R+D labs was the report's principal author). The study follows up on last year's publication of the EPA's report to Congress (pdf) on data center energy consumption. That report, among other things, estimated the range of savings that data center operators could achieve with varying degrees of technology and practice improvement. This more recent report is based on real world studies and was intended to validate the estimates in the EPA report.

Both reports are good reads if you are interested in reducing the megawatts being consumed in your organization's silicon (though the EPA report has been criticized as being a bit toothless). However, I should warn you that they are fairly long and detailed so the bedside table might not be the best home for them if you want to get through them, at least until the manga versions are released.

The EPA study estimated that "state of the art" technology and processes in the data center might cut energy usage by 55%, the more readily achievable "best practices" come in at 45% savings. State of the art includes a range of approaches including better server utilization through virtualization, better cooling techniques, improved power distribution, sensor networks, etc.

The more recent study, testing those techniques in working data centers, validates the EPA's estimates but also offers the initially surprising conclusion that legacy data centers can be retrofitted to achieve efficiencies close to that of new builds. That conclusion follows from the less surprising finding that the most bang for the buck comes from improvements on the "IT" side of the energy draw (energy efficient servers, virtualization, etc.) rather than from the harder to retrofit "site" side (cooling systems etc.). The dog wags the tail after all and if you can reduce the direct power consumption by the IT equipment, you will simultaneously reduce associated cooling costs whether in an old building with relatively inefficient HVAC or a shiny new one.

The last finding that I'll mention here is that it doesn't look like the time is right yet for widespread adoption of more advanced load management techniques outside of niche applications. The demonstration project had facilities that experimented with them, but the risk aversion that stems from high reliability requirements in production data centers has these experiments mostly restricted to centers that serve R+D rather than production functions.

Maybe one of the most interesting things about the report is what it doesn't (can't) say.

(continue reading)

The Last HOPE

I made the trek to a steamy hot NYC this weekend to attend one day of the three day Last HOPE (Hackers on Planet Earth) conference at the Hotel Pennsylvania. There was too much going to adequately cover it here (or even take it all in), but a few things stood out.

Steve Rambam's eye opening talk on the death of privacy for example. For a solid three hours in front of a standing room only crowd he weaved back and forth between the Orwellian theme of how our privacy is being ripped from us by everyone from Google to Choicepoint and the theme that seemed even creepier to him, self contribution. Over and over he expressed disbelief at how willingly we post our personal details everywhere from Twitter to Facebook while thanking us all the while for making his job as a private investigator that much easier. What the marketers and government don't actively take, we actively give. Naturally I twittered the whole thing.

Cell phone tracking; artificial-intelligence-assisted reality mining; 3000 cameras per square mile in Manhattan; facial, activity, and even gait identification software; government outsourced investigative databases shielded from FOIA requests; UAV-based license plate scanners; beating anonymity by correlating multiple datasets; unanticipated database repurposing; and on and on... Finally I could twitter no more and left the venue hurriedly fashioning a tinfoil hat from a burger wrapper while consigning myself to doubling the dosage on my meds.

I will say this though, there was something deliciously ironic about standing in a room chock full of hackers all listening at rapt attention to a three hour chillingly dystopic harangue on privacy loss while nearly every single one of them was wearing an RFID tag around their necks. Even better, the tag was tracking their every move around the venue and was linked to a comprehensive self-contributed profile.

Moving beyond the privacy nightmare stuff, there was hardware hacking to be found everywhere at Last HOPE. Tables were covered with broken open electronic toys and electronic components and were surrounded by hackers with smoking soldering irons.

Of the completed projects on display, one of my favorites was a something of a hybrid that projected a 3D image onto carefully placed strings.

(continue reading)

User Mediated Trans-Enterprise-Web Mashups?

There has been an on again off again discussion behind the scenes at Radar about the nature of the enterprise vs. the web and how they are defined not just by their technologies, but by their frames of reference. For my part, I think the enterprise view is defined implicitly by a planning mindset and a perceived scope of control that ends at the enterprise boundary. Whereas the web is too large for effective control so it tends to be an environment more conducive to serendipity and emergent behaviors. The web and the enterprise also differ in obvious cultural ways. Web culture tends toward speed and "good enough" while enterprise culture is informed by enterprise concerns like mission criticality, legal frameworks such as HIPPA and Sarbox, security, transaction volumes, and the like. These thoughts were still rattling around in my head last month when I arrived in Montreal for the weekend.

Just like every other year, as soon as I cleared customs I skipped the crappy exchange rates at the arrival area exchange vendors and headed upstairs to my favorite ATM machine in the departure hall. I needed to get cash for my cab ride to the center of town, only this time, the machine spit my card back out like day old sushi. I tried another ATM further down the hall with the same result.

After ten frustrating minutes of IVR traversal and the international roaming fees that went with it, I was talking to a Wachovia Bank customer service representative who politely suggested "you should have called us before you left the country, then we wouldn't have automatically blocked your card."

Apparently Wachovia (like many other banks) has decided the best way to reduce their risk of fraudulent transactions is to convert that risk to customer hassle with an algorithm that looks something like: IF Loc <> Home Country/County/City SET CardStatus to Blocked. My bank is now my mom and I have to call it and get permission before I am allowed to go out and play.

The funny thing about all this is that even though Wachovia suspected I hadn't accompanied my ATM card to Montreal, plenty of others knew that I had, including at least: AT+T (my cell phone provider), Verizon (my blackberry provider), Dopplr, USAA (I booked my flight with their credit card), Travelocity (where I booked the flight), US Airways, Plazes, Yahoo Fire Eagle (fed from Plazes and Dopplr), and naturally, the U.S. and Canadian Border Authorities.

(continue reading)

An ESB for the Web?

I spend a great deal of my time encouraging "enterprise people" to think more like "web people." Focus on adoption, use platforms to enable emergent capability, build the "generative enterprise," and that sort of thing.

So, imagine my surprise when I saw the web acting a bit like the enterprise with the launch of Gnip.

As the web moves toward a network of widespread transactional API's, each with it's own vocabulary, it is starting to look a lot like a legacy enterprise writ large or maybe like an industry eco-system. So we shouldn't be surprised to see web developers turning to solutions that their enterprise colleagues would find familiar.

Anyone who has spent more than five minutes in the enterprise world talking about SOA in the last five years (or spent time building "trading platforms" for industry consortiums prior to that) has probably drawn a picture on a whiteboard that looks something like this (see, almost identical):

Whether you have integrated line of business applications inside the enterprise or connected trading partners within an industry, that N squared connection problem will resonate with your experience. Webs of poorly documented point-to-point integrations are expensive to build, expensive to maintain, and impossibly brittle when the business changes.

And now the N squared problem seems like it might be beginning to resonate with web developers too now that they have to integrate to an ever growing population of API's. Plus, on the web, the additional limitations of a port 80 based infrastructure add to the nightmare by throwing the expense of constant API polling into the mix.

So, what to do?

(continue reading)

From the July NY Tech Meetup

I attended the New York Tech meetup last night with about 400 others in IAC's lobby on West 18th Street. Sort of like an Ignite but without the auto slide advance and proximate cash bar. Seven individuals or teams talked about their projects for five minutes each.

The lead off talk was from Transclick. Founded by an ex hedge fund manager (hedge fund managers leaving to do tech startups, things that make you go hmmm), Transclick does real time translation among 16 languages for IM, SMS, email for mobile devices. They are also beginning to leverage voice recognition tied to the translation engine. They've been around for a while and pretty widely discussed and feted so no need to add much more here. However, one interesting note came up during the Q+A. When asked about Twitter, Robert Levin, Transclick's CEO, claimed to be in negotiations to add language translation to the micro blogging service. Of course that will be great, as now native speaking spammers from all over the world will be able to follow me with less effort.

A few other quick mentions...

Pluribo is using natural language processing to summarize user reviews on Amazon.com. Delivered for now as a Firefox plugin, it analyzes a stream of user comments to find key words that relate to user concerns and automagically create a brief summary. Hovering over the key words in the summary brings up nice visuals that describe overall customer sentiment or issues to pay attention to for the item.

Daily Lit is either really great or mildly depressing depending on your point of view. Delivering books in bits and pieces via email or RSS, it is designed to fit literature into our harried lives through the channels we are already paying attention to. I can't decide if that is really cool or if it's like occasionally dropping a pearl into a Skinner box. Pellet, pellet, pellet, omg!!, Vronksy shot himself, pellet, pellet, pellet...

Cause Caller is a great example of the web's generativity in action. Fred Benenson's NYU thesis, it combines Media Wiki, Asterisk, and EC2 so that an individual citizen can describe a cause, link it to the politicians that might most readily influence the outcome, publish a call script, and automate the dialing. The most active cause on the site today is "say no to Telecom Immunity."

If you'd like to take a look at the rest of the projects that presented, they are Wakozi (a NYC-based delivery service), Cloudsmith (a cloud-spanning distro mapper), and Independence Year ("a workflow engine for taking the country back").

Philly's First Ignite was a Smash

At the risk of exposing my provincial roots, and now that the videos are up, I'd like to throw out some props to Philadelphia for last Wednesday's Ignite. I attended with a bit of trepidation and didn’t know quite what to expect but I was really blown away. While Boston has MIT, SF has Berkeley and Seattle has, well, Starbucks, Philadelphia proved it had a really fantastic mix of arts and socially oriented projects / talks that were uniquely Philadelphian.

Without a hint of Philly’s perennial shoulder chip, the mood in Johnny Brenda’s was two parts giddy, one part sweaty as we listened to talks from Philly Bike Share, the 100K House, Art and Poetry from UArts, Fab@Home, Etsy, The Food Trust, Indy Hall Worksharing, Jeff Stockbridge and his wonderful photographs from inside abandoned industrial era homes, No Carrier talking about Chip Music, Blake Jennelle’s talk about Philly as a vibrant locale for startups, an oddly prescriptive but delightful talk from Rick Banister about "establishing your personal aesthetic", and too many others to list... It was absolutely not what I expected and thoroughly wonderful. Attendance was over 300 and probably 50 or more were turned away from the perfectly gritty Fishtown venue. I was turned away too, but I’m pretty resourceful.

UArts is a real force in Philadelphia's emergent arts scene and it made a significant contribution to the evening both in participation and attendance (as an aside, they also provide space for Make:Philly's monthly meetings).

Thanks to Geoff DiMasi, Alex Gilbert, Vanja Buvac and Far McKon for putting it all together. Bravo Zulu.

Satan is on My Friends List...

Today I was invited to a Black Hat Briefings webcast that included the intriguing topic "Satan is on My Friends List - Attacking Social Networks." Naturally I registered; but that's not what this post is about. That title just sort of got me thinking about contrasting worldviews and how they relate to those frames Nat mentioned in his post on the enterprise the other day.

While both O'Reilly and Black Hat conferences attract alpha hackers, their editorial points of view tend to occupy opposite ends of some kind of Internet worldview continuum. Whether from the influence of their founding personalities or some kind of Sebastopol vs. Las Vegas center of gravity, they both represent what is possible but they tend toward different visions of that possibility. Maybe I'm reading too much into this, but for what its worth, I've attended both Black Hat and O'Reilly conferences and can't recall Satan making a single appearance in an O'Reilly conference program.

(continue reading)

Ignite Philly Tomorrow Night

I'm looking forward to Philadelphia's first Ignite event tomorrow night (the 11th) at Johnny Brenda's on Frankford Ave. I won't bother repeating all the details that you can find here but the doors open at 6:00 and a great lineup of talks starts at 7:00. You can RSVP on Facebook here if you'd like (it's not required). However, based on the number of RSVP's so far, I'd suggest getting there early. Hope to see you there.

Special Purpose Computing Focuses on Energy Efficiency

To improve the climate models that predict global warming, climatologists are seeking model resolutions on the order of 1 km. Unfortunately, building the required 200 petaflop machine with today's commodity-hardware approach would cost $1B and would result in a staggering 40 megawatts of power consumption.

A group of researchers at Lawrence Berkeley National Labs, who must be aware of the irony inherent in using 40 continuous megawatts to better predict global warming, may be returning supercomputing to its specialized roots but along a new vector (yes, weak pun intended). In addition to the Cray-era focus on raw power, they are emphasizing energy efficient computation where floating point operations per watt is the key metric.

Their approach has already been described at EcoTech Daily and the lab's Research News so I'm just going to summarize it here. They are working on specialized hardware consisting of 20 million very low power embedded processors (of the sort used in iPods and cell phones) wired together with the specific climate calculations in mind. By trading flexibility for efficiency, the design should achieve a ten-fold improvement in the floating point operations per watt metric and the resulting 200 petaflop machine is predicted to require only 4MW of power and cost $75M to build.

Their motivations in their own words:

"What we have demonstrated is that in the exascale computing regime, it makes more sense to target machine design for specific applications," Wehner said. "It will be impractical from a cost and power perspective to build general-purpose machines like today's supercomputers."

Specialized problems are amenable to specialized solutions and scientific computation seems particularly suitable to this kind of approach. However, on the web and in corporate IT where computing is both more general and inefficiently deployed, the first wave of energy efficiency improvements are being addressed primarily through a combination of virtualization and incremental improvements in commodity chip design.

I don't think software carpooling will be the only game in town for long though. While virtualization and dynamic provisioning are facilitating better utilization of existing hardware, virtualization comes with a performance cost of its own and can be no better than the hardware it is running on. Once you get four passengers in a V-8 powered SUV further improvements have to come from changing driving habits and modifying the vehicle.

As virtualization initiatives pick the low hanging fruit, further gains will come from fundamental hardware improvements (which may include analogous specialization) in concert with "best efficiency" dispatching that targets optimal server utilization in a dynamic server pool. An interesting example of this kind of approach is described here (pdf).

As I've touched on before, beyond that, a "systems view" to optimize the whole data center as it operates under changing conditions and with heterogeneous hardware might come next. Returning one last time to the carpooling analogy, this would be like a smart traffic routing system that keeps each car-pooling hybrid moving at its most efficient speed. The end result might be an optimally-sized mixed pool of specialized and commodity hardware each dispatched to operate the data center holistically at its best unit of work per watt.

GSM Cracking: Coming Soon to a Computer Near You via a Web Service

A web service that will make it easy and inexpensive to crack the GSM A5/1 encryption protocol, quickly enough for a call that is still in progress, is slated to launch at the end of April. Living right at the intersection of open hardware, open source software, software as a service, and cryptography, the service will reduce the cost and effort of cracking GSM call encryption by at least an order of magnitude.

The service is being developed by members of the GSM Software Project and demonstrates just how much things have changed in the world since the GSM system was designed. Various approaches to cracking both A5/1 (the European standard) and A5/2 (the weaker US standard) have been available for some time but this one is unique in that it should be available to researchers and hackers at the end of April in hosted api form instead of pdf.

Back in 1997 this overview of the GSM system declared that "Enciphering is an option for the fairly paranoid, since the signal is already coded, interleaved, and transmitted in a TDMA manner, thus providing protection from all but the most persistent and dedicated eavesdroppers." After all, such a radio encoding scheme made the signals invisible to typical radio band scanners.

Today, however, the availability of the Universal Software Radio Peripheral (USRP), an open hardware software defined radio that sells for about $700, combined with work being done at GNU Radio project to codify the GSM waveform (also targeted for the end of this month), makes this once reasonable point of view seem quaint. Good encryption is now a must and it appears that A5 no longer qualifies.

With USRP and GNU Radio making the waveform and encrypted frames cheaply accessible and the A5 Hacking Project's service easily breaking A5/1, anyone will be able to make a cheap GSM scanner. Today neither the complexity of the waveform or the encryption in use is adequate to keep a GSM call private.

(continue reading)

Recent Posts

• Open Source "Social App Server" Might Crack Garden Walls? on March 25, 2008
• How Technology Almost Lost the War, but Should Do Better on March 21, 2008
• Given Enough Eyeballs - Art Meets Open Source in Philly on March 14, 2008