Consolidating Federal Data Centers Won’t Be Easy

Map of planned data center closings
Planned data center closings (from

Derrick Harris at GigaOm offers an enlightening look at why the Office of Management & Budget’s plan to close 800 of the federal government’s 2,094 data centers (map) isn’t likely to happen. The problem: Federal agencies run tons of legacy apps that are going to resist the push for consolidation and virtualization.

Federal IT is a mess. Much of the software ranges from old to ancient. On workstations, Windows XP and Office 2003 is often still the rule. May agencies run multiple, tiny data centers.

The budget situation is not making things any easier. Data center consolidation, done right, could save a lot of money in the long run. But it will cost up front and the squeeze on agency funding makes that sort of investment difficult.

Apple and its Non-Acquisitions

Over at Technologizer, Harry McCracken has an excellent rundown on the long history of rumored acquisitions by Apple that never came to pass. Some of these hypothetical deals made at least superficial sense, most didn’t.

Here’s one test to apply to any talk of an Apple deal: Margin dilution. When financial analysts talk of dilution, they are referring to how the costs of an acquisition will reduce the equity of existing shareholders. Apple’s enormous market cap means that is rarely going to be a huge problem, but a company like Apple is going to be very wary of any acquisition that would seriously erode its very healthy profit margins.

That’s one overwhelming reason why Apple would be very unlikely to give more than about 10 second’s’ consideration to the rumored purchase of  Barnes & Noble. On the roughest sort of pro forma calculation, an Apple-Barnes & Noble merger would have reduced Apple’s profit margin in the most recent fiscal year from 23.5% to 21.9%. At most companies, a point and a half of margin is something to kill for. Since B&N offers no technology that Apple wants (strategic investments can be viewed through a different lens than purely financial ones) and  would saddle Apple with a great deal of real estate that it has no use for, the deal makes no sense on any level.

Can a New Approach to Wireless Beat Shannon’s Law?

For the past 60 years, electrical engineers have understood the hard limits that physics imposes on the data capacity of any channel. The law, formulated by Claude Shannon of Bell Telephone Labs, says that the data capacity, in bits per second, is a function of the bandwidth, the signal strength, and the noise in the channel.

Shannon's law formula
Shannon's Law (Wikipedia)

No one has yet found a way to break Shannon’s law, but Rearden Companies, the brainchild of Steve Perlman, who is behind the bandwidth-bending OnLive online games service, claims to have found a way to cheat significantly. A paper by Perlman and Rearden Principal Scientist Antonio Forenza, describes a technology called Distributed-Input-Distributed-Output (DIDO).

Normally, when a wireless channel, such as a connection to a Wi-Fi access point, is shared by two or more users, each user can only get a fraction of the channel’s capacity. But DIDO allows each user to communicate, in theory, up to the full Shannon limit of the channel.

The explanation for just how this works is complicated, but the technology uses array of antennas to create a non-interfering path between an access point and a user. Shannon’s law applies not to a particular piece of bandwidth but to each channel. Traditionally, we have thought of the two as the same, but DIDO spearates the concept and allows the link between the access point and each user to function as an independent channel within the dame physical bandwidth.

Normally, I am deeply skeptical about claims of fundamental scientific breakthroughs. My skepticism is mitigated by what Perlman has already accomplished with OnLive, which moves gaming data across internet connections with an efficiency that no one thought possible. Clearly, this guy knows how to move bits.

Don’t expect to see DIDO deployed anytime in the very near future. It requires significant changes to network design.. including interposing a DIDO data center between an internet source and an access point to encode data as well as sophisticated new antennas.  But it does hold real promise to to reduce the growing crunch on our wireless airspace.

Do Students Hate Textbooks More than They Like Sex?

It’s no secret that students hate both buying bloated, overpriced textbooks and lugging those bricks around in their backpacks. But we didn’t know how much.

A new survey sponsored by Kno, Inc.–which, not coincidentally, is in the business of e-textbook software–found that 73% of college students would do something they otherwise wouldn’t consider, including giving up sex, if they never had to shlep another textbook.

You can take that finding with a grain of salt, but there’s little doubt that the movement to e-texts is hitting an inflection point. The Kno study, conducted by Kelton Research, also found, more believably, that 71% of students want their texts to go digital.

A big driver, of course, is the rapid adoption of tablets, particularly the iPad, which make excellent textbook readers. An early attempt by Amazon to promote the jumbo Kindle DX as a textbook reader fizzled, mostly because of the limitations of the monochrome, video-free device. But the iPad is so natural for the job that Kno abandoned plans to come out with its own hardware to focus on iPad software.

The most recent big development in electronic textbooks was the announcement by Amazon that to would be renting texts for as little as 30 days and for up to 80% less than the print edition price. The books will be available on all devices that support the Kindle reader,  though I suspect that reading a typical textbook on a phone screen will not be a happy experience.

Amazon has initial partnerships with John Wiley & Sons, Elsevier, and Taylor & Francis. That will limit the selection of text available this fall, though other big players such as McGraw-Hill and Pearson will certainly join if the initial efforts shows legs.

That calculus text pictured above? Single Variable Calculus by James Stewart (Brooks Cole) is one variant of a widely used introductory text that is not available in digital form. A hardcover copy is still going to set a student back more than $100 and create a 2 1/2 lb. lump in a backpack.



Americans Elect’s Online Nomination Is a Recipe for Trouble

Americans Elect, a group that promises a new way of nominating a candidate for President in 2012, lept into prominence in recent days with a ringing endorsement from New York Times columnist Thomas Friedman. I don’t propose to debate the merits of the group’s ideas, but I want to take a look at the practicality of the proposed online nominating process. It’s not going to work.

Americans Elect logo The information Americans Elect gives on its web site is very sketchy, but the basic idea is that any registered voter can become a “delegate” simply by signing up. It’s not clear how, or if, voter registration is verified. All I had to do when I signed up was give an email address and create a four-digit PIN code. It appeared to me to be trivial to create multiple accounts using different email addresses; I asked Americans Elect about this but have not yet received a reply. [See update below]

There’s a good reason why we generally don’t see online voting for anything more serious than American Idol winners. Running a clean and secure online election is very, very difficult. Under the conditions we generally expect of formal elections, both security and anonymity as close to absolute as we can make them, it may be impossible. Last year, a modest experiment with online voting for U.S. military personnel abroad had to be suspended when it proved hopelessly insecure.

Systems that have been tried for serious online voting  generally require the distribution of voting tokens–generally a one-time password of some sort–through a secure offline channel. Often this is done by sending the information via postal mail. Americans Elect doesn’t say whether it plans to use such a system, but it would be complicated and expensive for sort of multi-stage nominating process it plans to use.

There’s a big risk that the Americans Elect nominating process could be turned into a circus. Unless exacting measures are in place to protect the integrity of the voting, the system could very easily be gamed (as has happened with American Idol voting.) It will be interesting to see what state election officials have to say about this process, although, in general, parties are given great latitude in how they nominate candidates, and in the view of election boards, Americans Elect will be a party.

By the way, anyone thinking of clicking the Donate button on the American Elect web site should be careful to read the fine print. Contributions to the organization, like those to any party or candidate, are not tax deductible.


UPDATE: I received the following unsigned reply to my question about measures to prevent multiple registrations:

Nothing, really. This is definitely at least something of a problem, but the way I see it, if we end up with with a meaningful number of delegates, the people with clone accounts will only be gaining minimal advantage to change things.

That said, if we can find a way to prevent this without causing significant usage/convenience problems, I’m all for it.

Sorry, but that drastically underestimates the web’s potential for mischief or malice. If Americans Elect want this process to be taken seriously, they’ll have to do a lot better.

PlayBook Wins Government Security Certification

Under the right circumstances, a bug really can become a feature.

Research In Motion announced today that its BlackBerry PlayBook tablet has been certified under Federal Information Processing Standard 140-2, meaning it is approved for nonclassified communications by U.s. government agencies and contractors who must comply with the Federal Information Management Standards Act. It is the first tablet to win this certification.

This speedy approval appears to be a direct consequence of what nearly everyone considers the PlayBook’s greatest weakness. The tablet has no email, contacts, calendar, or task program of its own, but instead serves as a sort of dumb display for a paired BlackBerry handset. Potentially sensitive information is never stored on the PlayBook, making it easy for the tablet to piggyback on BlackBerry’s existing FIPS certification.

The approval does give the slow-selling PlayBook a leg up in the government market. Unfortunately for RIM, a tight budgets and the ponderous federal procurement process make an imminent flood of government orders unlikely.


A Librarian Looks at the JSTOR Download Case

A Tweet by Nilay Patel (@reckless) of Verge called my attention to an intelligent post giving a librarian’s perspective on the charges against Demand Progress’ Aaron Swartz. Swartz faces federal felony charges for allegedly using unauthorized access to the MIT network and the JSTOR academic journal archive to download millions of articles.

Nancy Sims, copyright librarian at the University of Minnesota Libraries, dispels what have become some common misconceptions about the case. First, she points out, there are no copyright charges involved. And she notes that all databases such as JSTOR have restrictions on downloads and that the violations alleged in this case are pretty straightforward.

One place where I disagree with Sims is that she, like many others, compares the case to that of Lori Drew, a woman who was prosecuted for computer fraud for bullying a young teen in violation of MySpace terms of service. Drew was convicted, but an appeals court later threw out the case.

A key difference (besides that fact that no individuals suffered direct or indirect harm in the JSTOR case while Drew’s actions may have contributed to the target’s suicide) is that the actions laid out in the indictment show a pattern of willful assault on the MIT network and the JSTOR database. It is by no means a simple case, as Swartz’ defenders have claimed, of “downloading too many articles” or even violating MIT or JSTOR terms of service.

One other point: Sims, along with many others, questions the proportionality of the government charging multiple felonies carrying a theoretical prison term of up to 35 years. Prosecutors customarily throw everything they have into an indictment. I’d be very surprised if this case does not end in a plea bargain with no jail time.

Why Apple Can’t Chase the Low End

In a post here earlier today, Ban Bajarin dismissed the frequent criticism of Apple for failing to serve the low end of the computer market. Ben focused on consumers’ willingness to perceive, and pay for, value in Apple’s relatively expensive products.

But in wondering why otherwise knowledgable people keep hammering Apple on this point, it’s worth considering just how the company’s business model is working. Everyone else in the PC business depends on selling enormous volumes of product at razor-thin margins. This has steadily driven the average selling price of PCs downward, though NPD data show that the average retail ASP in the U.S. has stabilized a bit at around $600. Apple has exactly one product close to that price point, the $599 bottom-of-the-line Mac mini. In a world of $500 to $700 notebooks, the entry point for a Mac is $999 and goes up quickly from there.

And what has the refusal to chase the mass market done to Apple? It absolutely owns the market for computers selling for more than $1,000. As a result, with about 10% of the U.S. market and less worldwide, it is grabbing the lion’s share of industry profits. Apple’s operating margin from all products in the most recent quarter was 32.8% compared to 5.7% for Hewlett-Packard’s Personal Systems Group. HP, with total revenues of $127 billion a year, has a market capitalization of $76 billion. Apple, with just over $100 billion in revenue, is valued by the market at $362 billion.

With numbers like that, it’s just silly to argue that Apple should be chasing the profitless low end of the market (or, for that matter, offering low-cost, lower-margin versions of the iPhone and iPad.) The history of the tech business is full of companies that won large market share by cutting margins to the bone, or sometimes further. Apple is in the sweetest of all possible spots, and it would be lunacy to change the business plan.


Demand Progress’ Aaron Swartz and the Rush to Judgment

When people talk about the internet encouraging a rush to judgment, they are usually referring to tendency to assume the guilt of anyone charged with a crime, say a Casey Anthony or Dominique Strauss-Kahn, before all the facts are known, let alone a before a jury has rendered a verdict. But sometimes it works the other way.

On July 19, a federal grand jury indicted Aaron Swartz, a founder of Reddit and the political activist site Demand Progress on a list of charges related to the download of nearly 5 million articles from the academic journal archive JSTOR. Swartz and Demand Progress were smart. Even before the news of the indictment had spread, the site carried a blog post with the headline “Federal Government Indicts Former Demand Progress Executive Director For Downloading Too Many Journal Articles.”

The post allowed Swartz to set the tone for the commentary that followed. It quoted Demand Progress Executive Director David Segal as saying the indictment is “like trying to put someone in jail for allegedly checking too many books out of the library.” Numerous posters rushed to Swartz’ defense on Twitter many quoting the Demand Progress headline or picking up on the “checking out too many books” theme. A post by Oliver Day (@zeroday), “There are apparently no more documents in JSTOR, aaron schwartz stole them *all*,”  was widely retweeted.

The problem with all this is that the grand jury indictment paints a very different picture of  Swartz’ activities, making it sound more like he backed up a truck to the library loading dock (of course, the physical book analogy fails because, as zeroday’s Tweet suggests, any number of people can download a JSTOR document with depriving others of access.) The indictment describes a running war as JSTOR and  MIT worked to cut off Swartz’ downloads and acuses him of physically breaking into a MIT wiring closet to gain access after MIT administrators had barred his computer from the network.

The indictment makes no reference to the status of relations between JSTOR and Swartz though in general, a felony prosecution is at the discretion of the prosecutor, not a victim. And the indictment itself makes it clear that MIT, including the MIT Campus Police, participated actively in the investigation. (Someone who wants to try the things described in the indictment might do well to target a network less intensely and competently monitored and logged than MIT’s.)

I have no opinion as the Swartz’ guilt or innocence. But the charges of wire fraud and theft are serious ones and cannot be dismissed as “downloading too many articles.” As for what might be the motivations for the downloads, I am sympathetic with those who argue that JSTOR represents an obsolete and repressive system of extremely expensive controlled access to scientific journals (and also an invaluable resource to researchers.) But the future of paid, print journals is a complex subject that must be addressed by means other than vigilantism. Some informed debate on that subject would be helpful.

NOTE: The original version of this post misspelled Swartz’ name as Schwartz.

Senate Copyright Bill: What’s Really Proposed

Nilay Patel at What’s My Next offers an enlightening look at proposed changes in copyright law (S. 978) that would create new criminal penalties for streaming infringing videos. The bill has set off a bit of a panic in the gamer community because of fears that it could apply to people who post recorded videos of game play to services such as YouTube.

Patel does an excellent job of showing why this isn’t really the case, and along the way provides a useful service in explaining the arcane art of reading proposed legislation.

Sony Experia Play: The Case for Diversity

There’s a profusion of Android phones on the market and they all have an awful lot in common. They’re almost all rectangular slabs with displays ranging from 3 1/2 to nearly 5 inches and differ mainly in the color of the case and how ronded its corners are. The biggest difference available: Some models feature slide-out keyboards (and thicker bodies) while some offer just an on-screen keyboard. You’d never know from this lot that Android offers manufacturers almost total flexibility in hardware design.

Sony Ericsson Experia PlayThen there’s the Sony Ericsson Experia Play. ($200 in the U.S. with two-year Verizon Wireless contract.) I don’t know whether the Play can revive the flagging fortunes of Sony Ericsson, but I’m glad to see themselves try something really different.

At first glance, the Play looks like every slider keyboard Android: Shiny black, rounded plastic case, 4″ 854×480 pixel display, 6.2 oz. (175 g), a bit over half an inch (16 mm) thick. But when you slide the bottom out, instead of a keyboard you see a game controller pad modeled on those used with the Sony PlayStation 3. In place of keys, you get a full set of gaming controls: the familiar directional and symbol pads, physical select, start, and menu buttons, and two circular areas that are designed to simulate analog joysticks. Games can use all these controls, plus the touchscreen and motion sensitivity.

I don’t play game a lot and I really can’t say how a hard-core game would feel about the Play, but I found it a lot of fun. The screen is crisp, though I had to turn up the brightness manually for satisfactory play. The Qualcomm Snapdragon processor provides more thjan adequate performance.

When not playing games, the Play becomes a standard Android phone with all the expected features. But it’s not likely to be anyone’s choice unless the gaming is important–then it’s just a thick handset without a physical keyboard.

I’d like to see more manufacturers experiment more freely with what Android has to offer. It’s the only viable phone operating system that offers this sort of freedom. Apple’s iOS, Research In Motion’s BlackBerry and (so far) Hewlett-Packard’s webOS are not available to other OEMs. Microsoft, burned by the fragmentation of design of Windows Mobile phones, is keeping a very tight rein on Windows Phone 7 designs.

Android creates the possibility for considerable variety combined with support for core Android features and apps. It’s a tricky balancing act, but it would be good to see more specialized handsets hit the market.


RIAA: One Bad Idea Deserves Another

You can always count on the folks at the Recording Industry Association of America to take a bad situation and make it worse.

For the past few weeks, the juvenile delinquents at LulzSec and Anonymous have been breaking into networks and web sites and bragging about their exploits to prove–we’ll it’s not clear just what they are trying to prove other than that they can do it. For the most part, the damage has not been terribly serious. It’s a bit like the heyday of graffiti, when the inability or unwillingness of authorities to stop spray-painting vandals created a pervasive sense of disorder in big cities.

Now the RIAA has come forward, arguing that the proper response to the outbreak of network vandalism is the passage of a truly bad law called the Protect IP Act. In a blog post, RIAA Executive Vice President Neil Turkewitz argues that the way to restore order is to give the government broad powers to block access to web sites that are accused of distributing pirated works. “And in a world where hackers set their sights on new targets every day – most recently the official United States Senate websiteallegedly the CIA’s public website and Arizona’s law enforcement database – do we think a lawless Internet defended to the extreme is a good thing?,” he writes.

The real problem with the LulzSec and Anonymous is that they are making the FBI, Secret Service, and other agencies charged with enforcing order on the internet look silly. Back in the 1970s, law enforcement tended to ignore graffiti because officials felt they had more important things to worry about. This was a mistake because the garish spray painting told the public the police could not do their job. A crackdown on internet vandals is in order, but we shouldn’t use this as an excuse for another bad law to save an industry from a failed business model.

A Sure Sign of Real Trouble at RIM

The senior Research In Motion executive who chose to vent his (or her) frustration in a open letter to Boy Genius Report may not have chosen the most graceful way to make those views known. But the writer may well have exhausted other means of communications. Certainly, RIM’s response suggests strongly that the increasingly troubled company’s leadership still isn’t hearing what it needs to hear.

The fact is that the open letter was an accurate analysis of the challenges facing RIM and was full of generally very good advice. The response is dismissive and described RIM’s current situation as a time when it is “necessary for the company to streamline its operations in order to allow it to grow its business profitably while pursuing newer strategic opportunities” after “a period of hyper growth.”

Streamlining and, above all, focus is exactly what the letter writer argued for. Mike Lazaridis and Jim Balsillie should give it another read with more open minds.

The Case For (and Against) a Cheap iPhone

There’s been a fair amount of buzz in the last few days about Apple introducing a cheaper iPhone this fall and in “The iPhone Is Too Expensive” at Slate, Farhad Manjoo makes a good case for Apple doing just that. But I seriously doubt that Apple will do so because, while the arguments for going downmarket make sense for any other manufacturer, that just isn’t how Apple works.

It seems to me that the surest way to go wrong in anticipating an Apple move, and I have done this often enough myself, is to assume that the company  gives a damn about market share. Apple is driven by margin and total profits, not by share, and this strategy has made it by far the most successful consumer electronics company in the world.

Yes, Apple could do a de-featured iPhone that could sell for $200-$250 without a contract and compete with a horde of generic Android handsets. It would undoubtedly increase Apple’s market share, especially if it was sold with prepaid service. All Apple would have to do is accept tiny margins and sell a product that the company knows isn’t as good as it could be. That just isn’t in Apple’s makeup.

Instead, I expect they will bring out a new iPhone in September (I’m guessing about the date, like everyone else) and keep the iPhone 4 in the lineup at a sharply reduced price. (A year after the introduction of the iPhone 4, you can buy an iPhone 3GS from the Apple store for $49. How much cheaper do you want it?)


Final Cut Pro X, Apple, and the Enterprise

Final cut Pro X iconA couple of days ago, I wrote about how Macs has become the overwhelming computer of choice for tech elites. No sooner had I done this than Apple offered glaring proof of its limitations as a provider of technology for professionals–or as a vendor to the enterprise.

Final Cut Pro X is the successor to Final Cut Pro, which has become the non-linear editing software of choice for professional videographers and filmmakers. (It also replaces Final Cut Express, a prosumer version.) The problem is that X is a completely new program, with new ways of doing things. It is incompatible with project files for older versions and lacks many features that pros have come to rely on.
Continue reading Final Cut Pro X, Apple, and the Enterprise

Macs and Windows: Why Tech Elites’ Choices Matter

Over the past couple of years, Windows laptops have been becoming rarer and rarer at events where tech reporters, bloggers, and analysts gather. Not so long ago, Windows PCs (including netbooks) outnumbered Macs at these affairs by two or three to one. Today, that ratio is at least reversed. The netbooks have all but disappeared and their place has been taken by tablets, nearly all of them iPads.

Apple has gained significant share in the laptop market, but not at anywhere near the rate of this shift. And this Mac dominance is a tech industry phenomenon. This week I was at a Ford Motor Co.-sponsored gathering of bloggers and magazine writers, most of who write about things other than tech. Windows PCs were dominant, though I did see plenty of iPads.

The overwhelming preference for Macs among tech elites has real consequences. There’s a reason why they are often called influencers: They have a lot of effect on other people’s choices. I try to be fair in everything I write, but it’s hard for me to work up much enthusiasm of anything Window-based these days. When asked for a recommendation, I always go with Apple unless there is something specific about the user’s requirements that argues for Windows. And that’s doubly true if I think I will end up supporting the purchase.

Why do tech elites prefer Macs? It’s certainly not because they love Apple, which regularly sets new standards for being hard to do business with. I think there are several reasons. One, oddly,has to do with price. The best argument against Macs is that you can buy a perfectly serviceable Windows notebook for around $500, while the entry price for a Mac is $1,000 or more. But the fact is that members of the tech elite tend to buy (or have employers who will buy) relatively high-end equipment. Spec-for-spec, Macs are not particularly more expensive than Windows systems, so the price differential is not an issue in this market.

Second, tech elites care, often passionately, about their technology and Apple equipment is a joy to use. And for people passionate about their technology, esthetics matter, and no one comes close to Apple. On the rare occasion when an HP or a Dell comes up with a really handsome product, it still must swim in a sea of cheap-looking  junk.

Apple hasn’t made an ugly product since it retired the eMac. And the 13″ MacBook Air on which I am writing this is, for my purposes, the best laptop I have ever used–by far. The combination of light weight, terrific battery life, and snappy performance (for the sort of light-duty work I do on this system) cannot be matched by anything else on the market. (If it had a 15″ display without being any bigger or heavier, an obvious impossibility, it would be perfect.)

Then, of course, there’s the software. The yawning gap that had opened between Mac OS X and Windows during the Vista fiasco has narrowed considerably but in a home or small business environment, Mac software is much easier to set up and maintain. Take setting up a networked printer. In Windows, despite improvements in Windows 7, this remains a black art, largely because the paradigm is designed for enterprises and IT administrators. With OS X, you connect your printer to the local network and you Mac finds it, using Apple’s dead simple Bonjour protocol. IT departments may see job security in complexity, but for those of us for whom maintain our own or other peoples’ systems is a distraction, simple is a huge advantage.

Microsoft’s business model depends on keeping large enterprise customers happy. The big PC makers’ business model depends on selling huge volumes of low-margin product. That means that neither can compete with Apple among customers who demand the computing equivalent of a Lexus or a BMW and are willing to pay for it.



Tech Patent Fights: What’s at Stake

Recent days have been filled with news about patent disputes. Lodsys, a company that claims fundamental patents on in-application purchases, fired off another batch of suits against alleged infringers. Apple and Nokia resolved a complex legal fight over smartphone patents. Dolby Labs sued Research In Motion. And the U.S. Supreme Court told Microsoft to pay up on a judgment that technology in Office infringed on a patent held by tiny i4i LP. Continue reading Tech Patent Fights: What’s at Stake