» Archive for June, 2012

Warning: The Cloud Isn’t All It’s Cracked Up To Be

Saturday, June 30th, 2012 by Jonathan Spira

This is not the first time I am writing this column.

Cloudy forecast for cloud computing?

While I last tackled the topic of Amazon’s outages some 14 months ago, when Amazon’s cloud-based data center service went down in a big way, Amazon has had several more highly-publicized outages, most recenty in April and in the middle of June.

These outages impact businesses and consumers alike.  Home movie fans who tried to use Netflix last Friday were disappointed.  Instagram users were in shock because they couldn’t share photos.  And Amazon has thousands of other customers, many of whom found themselves in the same boat.

This time, the problems were caused by the severe East Coast storms that left over two million people without power, but the cause doesn’t matter.  It’s the lack of preparedness for dealing with storms and outages that worries me.

If you’re thinking, “oh, but this is in the cloud,” I have some news for you.  The cloud has to have an earthly connection somewhere and redundancy doesn’t seem to be Amazon’s strong suit (or that of its customers).  The failure occurred at an Amazon location in Virginia, and that’s where many of these companies had their data.  They didn’t seem to think it made sense to put it in a second place, perhaps for safe keeping or in case a storm blew in and knocked one location off line.

Amazon did do a good job of updating its status messages with somewhat terse language (example: we are “investigating elevated error rates impacting a limited number” of customers) but that didn’t bring the data center back online any faster.

The problems weren’t over Saturday morning although the company said that some of its servers and services were back online.  ”We are continuing our recovery efforts for the remaining EC2 instances,” the company posted shortly before noon.

If you were an Amazon customer, these messages were all you would get.  Presumably your websites were offline as were any services, so you had to hope that Twitter wasn’t using Amazon’s cloud in Virginia that day as that’s how many Amazon customers were telling their own customers about the outage.

The point of my bringing this up is simple and something I’ve said before, numerous times: more and more data storage and processing have been off-loaded into the cloud without the appropriate precautions being taken to ensure data accessibility and redundancy.

I’m sure that we’ll see people questioning the viability of cloud-based services and storage because of this occurrence – but taking steps to ensure that a single outage doesn’t take thousands of companies down in one fell swoop seems to be quickly forgotten until the next time it happens.

It’s not up to Amazon. If you’re an Amazon customer, you’re clearly on your own – and you need to be working on a plan to ensure uptime.  If not, your customers will, by going somewhere else.

Jonathan B. Spira is CEO and Chief Analyst at Basex and author of Overload! How Too Much Information Is Hazardous To Your Organization.

Google X Lab Hunts YouTube for Cats

Friday, June 29th, 2012 by Cody Burke
feral cat in Virginia

That computer is watching me...

When left to its own devices, what does one of the largest neural networks for machine learning in the world use its 16,000 computer processors and one billion connections do? Solve complex environmental problems? Crunch scientific data to illuminate the mysteries of deep space? No and no. Turns out, the huge, powerful neural network taught itself to recognize cats (and humans) in only three days.

Researchers from Google and Stanford working at Google’s secretive X Lab connected 16,000 processors (still far short of a human’s estimated 80 billion neurons) and fed the neural network digital images extracted at random from 10 million YouTube videos. The machine was then left alone to learn what it could with no instruction on how to proceed. For three days, the network pored through the images, making connections and finding commonalities between objects. Next, the researchers attempted to see what the computer could identify from a list of 20,000 items.

By learning from the most commonly occurring images the computer was able to achieve an 81.7% accuracy rating in identifying human faces, 76.7% accuracy at identifying human body parts, and 74.8 percent accuracy when identifying cats. The increases represent a 70% jump in accuracy compared to previous studies.

The network constructed a rough image of what a cat would look like by extracting general features as it was exposed to the 10 million images, in much the same way that a human brain uses repeated firing of specific neurons in the visual cortex to train itself to recognize a particular face. What the experiment proved is that it is possible for the computer to learn what something is without it being labeled; the neural network created the concept of both humans and cats without being prompted.

Machine learning lies in the use of algorithms to allow computers to evolve behavior based on data by recognizing complex problems and making intelligent, data-based decisions. The problem is that when faced with complex problems with large data sets, it is next to impossible for all possible variables to be covered, so the system must generalize from the data it has. In the case of the cats, Google’s neural network was able to generalize what humans and cats look like with relatively high accuracy.

The potential applications of this kind of neural network are broad. Speech and facial recognition, as well as translation software would benefit from machine learning that only requires vast amounts of data with no hints or guidance from human operators. The current Big Data movement is providing huge quantities of data, and it is encouraging to know that there may be actual applications for it.

Reflecting the confidence that Google has in the project, the company is moving the neural network research out of X Lab and into its division charged with search and related services. Expect to see larger neural networks with even higher accuracy rates constructed in the near future. Hopefully we can move on to using them for something more important… perhaps even recognizing dogs?

Cody Burke is a senior analyst at Basex. He can be reached at cburke@basex.com

(Photo: Stavrolo)

Microsoft Gets in the Tablet Game

Thursday, June 21st, 2012 by Cody Burke
Microsoft Surface

What am I? Tablet or laptop?

Against the backdrop of continued Apple dominance of the tablet market, Microsoft has thrown its hat in the ring with the Microsoft Surface. The new tablet comes in two models, a slimmed down version running Windows RT, (the company’s simplified mobile Windows for tablets) and a professional offering that runs full Windows 8. Both enter a crowded but lopsided tablet market where the iPad dominates the full size tablet space and a confusing plethora of Android tablets fight for the scraps.

Taking cues from Apple, Microsoft staged the announcement with much secrecy leading up to the event, and by all accounts, put on a good show for the eager crowd. This is Microsoft’s first attempt to build its own tablet hardware; the company is currently relying on third party hardware developers to release multiple upcoming Windows 8 tablets. Microsoft’s last attempt to take over hardware design on its own and do battle with Apple was the ill-fated Zune. The Zune, despite getting positive reviews for its design and functionality, never gained enough marketshare and now only lives on as software.

Although the new Surface tablets looked good in the demos, there are many questions remaining about the devices. What we do know is that both models feature a 10.6”, 16:9 “ClearType” 1920 x 1080 display, front and rear-facing cameras, an SD slot, USB port, magnesium casing, Gorilla Glass, and a kickstand for landscape view. Wi-Fi is enabled via a 2×2 MIMO antenna. The Surface for Windows RT is 9.3 mm thick (the iPad is 9.4 mm), weighs .68 kg, and runs on an ARM processor. It runs Metro apps, so it will deliver a mobile experience comparable to competing Android tablets and the iPad.

The Surface also features a pressure-sensitive keyboard hidden in the tablet’s cover, called Type Cover. When closed, the cover looks similar to the iPad’s Smart Cover, but when open, a full keyboard is revealed on the inside. Using the tablet’s kickstand allows the screen to be propped up in landscape orientation with the keyboard rolled out in front, approximating the experience of working on an Ultrabook or netbook.

Most intriguing is the Surface for Windows 8 Pro, which runs full Windows 8 on an Intel Ivy Bridge Core i5 x86 processor. It is slightly thicker at 13.5 mm, adds a USB 3.0 port, and is available with either 64 or 128 GB of storage. The critical point about the professional Surface tablet is that it is essentially a full-featured PC in a tablet form factor. When combined with the Type Cover, the Surface is almost bypassing the iPad to go up against the MacBook Air and other Ultrabooks (for better or worse). If the typing experience is good, that is a win for Microsoft as it adds functionality that the iPad currently lacks and allows it to attempt to woo mobile workers. If Type Cover turns out to be gimmicky and not suited for real work, then the Surface will remain in the tablet arena.

Running a full version of Windows 8 allows the device to run Office, Photoshop, or any other professional desktop application. Not having to depend on mobile apps could be a game changer, particularly since Windows 8 has been designed from the ground up as a touch-screen friendly iteration of Windows. Past Windows tablets have struggled with trying to squeeze the desktop metaphor onto a touchscreen tablet, with poor results.

The question now is who will actually buy the Surface. The base model may struggle to compete with Android tablets and the iPad unless it truly brings something new to the tablet experience (I’m not sure the fold out keyboard will be enough). The professional model however, if it does indeed provide a full Windows 8 workspace and the ability to run desktop applications, may be able to differentiate itself from the iPad in ways that the BlackBerry Playbook, HP TouchPad, and the legions of Android tablets have not been able to.

Regardless, the Surface is a significant and bold move for Microsoft, and you have to respect the audacity of going up against the dominant iPad, and maybe, just maybe, coming out ahead on some features. For Microsoft’s (and the consumers’) sake, lets hope that the Surface is not a repeat of the Zune.

Cody Burke is a senior analyst at Basex. He can be reached at cburke@basex.com

Is Information Overload Yesterday’s News?

Monday, June 18th, 2012 by Jonathan Spira
London Gazette

No more headlines for Information Overload?

Information Overload continues to be a costly problem but it also appears to be losing some degree of mindshare.

This is not completely unexpected.  Many such areas – knowledge management comes to mind – go through phases.  At some point, knowledge management was viewed as the savior of the Information Age and at another point, it was seen more as a pariah.

Google Insights, a new tool currently in beta, helped shed some light on where we stand with Information Overload today.

Following a pattern set forth by Thomas Kuhn (albeit for a paradigm shift, not a technology trend) in his groundbreaking work, The Structure of Scientific Revolutions, a revolution occurs in which an anomaly cannot be explained by current theories and the current worldview.  This throws the world into a form of crisis and turmoil and an intellectual battle occurs between the adherents of the old paradigm and the proponents of the new one.

Eventually, the new worldview takes hold.  Kuhn also maintained that the proponents of the new paradigm cannot be those who were adherents of the old one. They simply had to die out and, once that occurred, a new paradigm was firmly in place and the crisis was over.

(A good example of this is the period of turmoil in science as the Einsteinian Relativistic worldview supplanted the Maxwellian electromagnetic paradigm.)

Is a scientific revolution afoot in the world of Information Overload?  That is something that can only be determined in hindsight and not while in the midst of the transition, but there are a few signs that something is changing dramatically.

One way of looking at a worldview on a particular topic today is by analyzing search queries.  In particular, Google’s new Insights search analysis tool provides a visual representation of regional and worldwide interest in a particular search term.  Google Insights shows the relative number of searches on a query, with the number normalized and scaled to a value from 0-100.  It is important to note here that the tool does not provide the exact number of searches, just the relative number.

When I applied Google Insights to the term “Information Overload,” I found something rather unexpected.

Despite the abundance of news articles, books, and technology solutions addressing the problem, the general public’s interest in Information Overload has steadily been in decline since at least 2004 (the first year for which data in Google’s Insights is available).  In 2004, the relative search number for Information Overload was 100 on a 0-100 scale.  The number of searches declined steadily from 2004 to June of 2006, when the search rate leveled out at 20, where it roughly remains today.

A small side note here; if you look in Google Insights for Information Overload without quotes, the top search term that it finds is actually “overload of information.”  When quotes are used (“Information Overload”), the top search becomes “internet information overload” and “definition information overload.”

Globally, Australia is the country most concerned with the problem and leads with the highest search rate at 100, with the U.S. (81), and the U.K. (78) following.  Germany (number seven on the list) seems very unconcerned, with a search rate of only 31.  Within the U.S., Maryland (search rate of 100), the District of Columbia (90), and Washington State (88) are the top three locales where people have  searched on the term Information Overload.  However, when broken down by city, New York is by far the largest source of searches on Information Overload with a search rate of 100.

Contrast this with a trend that has recently garnered much attention, “Big Data.”  If you use the Insights tool to search for both Information Overload and Big Data (in separate searches), things really start to get interesting.

Big Data had been seemingly absent from search queries (hovering at a relative score of roughly three) until around September 2010.  At that point, Big Data jumped from five to 38 by the end of 2011.  Big Data’s search rate continues to grow and currently sits at 100 while Information Overload’s rate has remained essentially unchanged since the spring of 2007, hovering between 20 and 30.

While Google’s Insights tool gives us interesting data, it is limited to presenting relative search queries, not exact search numbers, making any direct comparison difficult.  What is clear is that Big Data is a term that has rapidly gained mindshare over the last few years, while Information Overload has somewhat fallen out of favor.

Jonathan B. Spira is CEO and Chief Analyst at Basex and author of Overload! How Too Much Information Is Hazardous To Your Organization.

Music and Productivity: All About Context

Thursday, June 7th, 2012 by Cody Burke

Must find the right song for the task at hand...

Last week Jonathan Spira explored the possibility of music being a distracting influence on knowledge workers. Although there is some truth to that, music is a topic that engenders such passion that getting knowledge workers (or anyone) to give it up would almost surely incite rebellion. A more nuanced approach is to find what works for the specific individual.

Music’s relationship to productivity is extremely complex and varies based on context. Factors such as the specific task being performed, individual preference of the knowledge worker, and the tempo and genre of the music make a huge difference.

To start with, research conducted in 2005 by the University of Windsor in Canada concluded that any increase in productivity from music may be related to the positive feelings that individuals get from that music. Essentially, people who listen to music that they like, feel better about themselves, and thus do better work. Forcing people to listen to music they don’t like, or forcing them to not listen to music at all could have the opposite result. The critical point is that people feel better when they listen to music they like.

The tempo and genre of the music is extremely relevant to any discussion of the impact of music on productivity. In my former career, I was a professional chef. In busy restaurant kitchens, I found it is essential to have fast-paced, aggressive music playing, preferably at loud volumes. Electronic dance music, heavy metal, and hip hop all worked well. The music would set the pace for the cooks and provide just the right amount of distraction that allowed everyone to relax, let muscle memory take over, and fall into a “zone.”

Although the music worked for us, it didn’t always work for others. Frequently, waiters would complain that they found the loud, fast-paced music distracting. Although perhaps this was simply a matter of personal taste, it also is important to consider that they were doing a completely different job – and using a different part of their brains. They had to remember orders, talk to customers, and generally be attentive to other human beings. We cooks needed to put our heads down and dig our way out of a rapidly growing hole by performing practiced, repetitive motions within a tightly knit team, often for hours on end. Different job requirements equal different musical needs.

In my current life as a writer and technology analyst, I work for the most part without music. It helps being in my home office, so I do not have to try to drown out background noise from co-workers, but for the most part I write and think more effectively with only ambient house noise. However, when I am faced with a repetitive task under a deadline, I revert to my cooking days and play mainly electronic dance music, although not as loudly as I used to. I’m getting old, I suppose. The beat sets a nice pace and keeps my brain moving quickly.

Context matters, and there is no one-size fits all answer when pondering how music impacts productivity. Find what works for you and enjoy it.

Cody Burke is a senior analyst at Basex. He can be reached at cburke@basex.com

Music: The Great Distraction?

Friday, June 1st, 2012 by Jonathan Spira
Vienne State Opera

Is even the Vienna State Opera a distraction?

Recently, I noticed more and more knowledge workers wearing headsets while at their desks. Not just earbuds, but noise-cancelling headsets, the kind you’d expect to see on an airplane (indeed, Bose invented the noise-cancelling headset over 20 years ago specifically for use by pilots).

While some people are simply using the active (or in some cases passive) noise-cancelling feature, most are listening to music.

What choice they make in music will unquestionably impact their productivity.

Years ago, when we planned the first Basex office space large enough to require a PA system and speakers, we also installed a CD-changer (these were pre-MP3 days), which played music in the background.

The type of music was subject to great debate but I insisted on music without any lyrics and without any particularly distracting beat. This translated into classical and jazz and, to be honest, was not a popular decision in the office.

But I had my reasons, the primary one being that listening to music with lyrics is a distraction. We had yet to start our research on distractions and interruptions (the timeframe I am writing about is ca. 1996) but, to paraphrase Supreme Court Justice Potter Stewart, I know a distraction when I see it.

We actually did some informal tests by playing music from various radio stations and I quizzed some of my colleagues after different types of music had been playing. What I found was quite revealing: favorite songs pretty much caused work to stop, DJ chatter slowed things down, and classical music was soothing.

This is because the lyrics or banter were essentially overloading the brain and it had to work overtime to keep the focus on work versus the music (this is an activity that takes place in the brain’s prefrontal cortex). As we have learnt since then, distractions take a significant toll on workplace productivity and each interruption comes with the penalty of what I named “recovery time,” i.e. the time it takes the knowledge worker to return to the task at hand after the interruption. (For an in-depth discussion of recovery time, the reader is referred to Information Overload – Is there a silver bullet?)

Recently, I read of several studies that took place in Japan on the topic. One linked music with lyrics to lower levels of concentration. Another found that workers without strong feelings about music found music less distracting than those who either really liked or disliked it. A third study, this time in the U.S., found that listening to hip-hop music resulted in a significant reduction of reading-test scores.

So back to the original topic of headphones for a moment. While having a choice in music may be something workers today feel is an entitlement – after all, thanks to the Internet, when it comes to music, the world is your oyster – there may be a significant downside when it comes to productivity depending on what each individual chooses to listen to. (In the interest of fairness, there may be an upside as well but that is beyond the scope of this analysis in which I am focusing on how certain types of music can overtax the brain while knowledge workers are attempting to work.)

As for me, I listen to Radio Swiss Classic on the Internet. Since it’s an Internet-only radio station (run by a major broadcaster, SRG SSR idée Suisse), there are no commercials and no news reports, just a very brief announcement with details on the work being played.

Jonathan B. Spira is CEO and Chief Analyst at Basex and author of Overload! How Too Much Information Is Hazardous To Your Organization.