Here's the filing, which is sure to put CEO Carl Russo back in the spotlight. Besides being one of the funniest, most free-wheeling executives one is likely to meet, he's also famous for having sold start-up Cerent Inc. to Cisco Systems in 1999 for $7 billion. Cerent had already filed for its IPO before Cisco came calling, paying one of the frothiest valuations of the late 1990s.

But while Russo may have been the front-man for Cerent and now for Calix, the founder of both companies was Mike Hatfield (He actually had a few co-founders on Cyan). While less well-known, he's got a sterling record when it comes to knowing where the puck is going in networking. After stints at former high-fliers DSC and AFC during the formative days of the Information Superhighway, he created Cerent in 1997 to build optical networking gear used by phone companies to move massive amounts of voice and data traffic through the "core" of the Internet that connects major cities and regions.

Of course, demand for networking gear of all types crashed just a year later with the Net Bust, as it became clear that far too much capacity had been built. Nonetheless, Hatfield founded Calix just months later to take a crack at what had long been considered the least appealing part of the networking market: the so-called "access" gear that delivers traffic over "the last mile," from the phone companies' central offices located in towns and neighborhoods to each subscriber's home. Traditionally, companies competed hard to make a thin profit on this more commoditized gear. But Hatfield sensed that Calix could attract a lot of attention if its boxes helped phone companies deliver more than just phone service. After all, cable companies were trying to add phone service to their menu of offerings, and Calix' gear would let phone companies add TV and other services to their basic voice plans. According to its IPO filing, the company had $250 million in revenuees in 2008, though it lost $12.9 million on the year (so evidently Calix hasn't been as successful at figuring out how to fatten those margins on access gear.)

So if Cerent was about the "core" and Calix was about the "edge," guess where Cyan is aiming? That's right--at "the middle mile." This is gear that connects the central offices to the core network. While most venture capitalists are still loath to invest in any networking start-ups (Hatfield figures they lost $2 billion on them during the Net Bust), he figures this section will be the next big chokepoint on the Internet. The reason: millions of people are now routinely consuming and sending high-definition video and other weighty digital fare, to a broader range of devices, such as the iPhone. But much of this middle mile still contains huge amounts of old copper lines rather than higher-capacity fiber-optics. Cyan's gear is designed to help carriers make the most of what they have, as they move to networks more capable of handling the load.

More than technology, Hatfield says Cyan is one of a new generation of networking companies with much lower-cost business models. That's critical, because carriers can't afford to just keep buying more of the same pricier gear as Net traffic continues to soar. While it has typically cost hundreds of millions of dollars to build a cutting-edge networking company in the past, he thinks Cyan will cost a tenth as much. It's not just that consumers are demanding so much more bandwidth, but that they're not willing to pay for it. "Today's bandwidth hogs are the consumers of the future," he says."Consumers aren't going to pay ten times the money for ten times the bandwidth," he says. See a video interview with Hatfield, with the site Light Reading, here.

He says Cyan's costs are so much lower because it relies on open source software and off-the-shelf communications chips--rather than proprietary code and chips (sounds like Arista, which I wrote about in the magazine a few weeks back). Evidently, Cyan is off to a promising start. Hatfield says the company already has more than 20 customers, and more are on the way thanks to the Federal broadband stimulus program.

Neither of these partners may have the funds for such a deal, however. Clearwire has just raised more than $1.5 billion in funding; but that's less than half of what it needs for its own network build-out through 2013, according to analysts. MetroPCS and Leap are still small companies, struggling to keep growing amidst rising competition. MetroPCS's subscriber additions in the third quarter were less than a third of what they were a year ago.

Yes, it is possible that all these struggling, smaller competitors will decide to band together and to fund one network, to be used by all -- say, Clearwire's. Most of them address a different market segment, so they won't compete with each other too much: T-Mobile goes after the hip, young crowd (though it's also pursuing prepaid customers). Clearwire offers mobile broadband services for laptops. MetroPCS and Leap have made their names on prepaid wireless plans.

But I would argue that what T-Mobile USA needs is to be paired up with a cash-rich, well-to-do giant, instead. After all, you put a bunch of struggling companies together, and you often end up with a large struggling company.

It seems to me that DT should, instead, look in a different direction -- to AT&T, for example. AT&T is healthy and has the funds to help T-Mobile out. It also currently uses the same type of networking equipment as T-Mobile, and could help T-Mobile migrate to next-generation technology more smoothly. While such an alliance could, potentially, raise the anti-trust flag, a deal could, perhaps, be structured in such a way as to overcome such concerns.

In its Nov. 19 earnings report, Dell said sales fell 15% to $12.9 billion. Net income fell by 54% to $337 million, or 23 cents per share after excluding certain one-time items. Wall Street analysts had expected Dell to earn 28 cents a share on sales of $13.1 billion. A year ago, Dell reported earnings of $727 million, or 37 cents per share, on $15.2 billion in sales.

Shares of Dell fell by nearly 6% in extended trading after the report. At the end of regular trading Nov. 19, Dell’s stock closed down 19 cents, or 1.2%, at 15.87.

The weakness was spread across nearly all of Dell’s businesses. Sales to large businesses bore the brunt of the declines as information technology departments continue to keep a tight rein on costs. Nearly 80% of Dell’s sales are to businesses and government customers. “We are losing share in the aggregate” because of a heavy reliance on commercial sales, Gladden told reporters during a conference call after the results were announced.

Dell didn’t see much benefit from Microsoft’s launch of its new Windows 7 operating system on Oct. 22, since Dell’s quarter ended eight days later. In the two weeks leading up to the launch, customers put off PC purchases to avoid buying machines with older software running on them, Gladden said. “We built a little backlog as a result, and we’ll ship through that in the fourth quarter,” he said.

Dell’s gross profit margin came in at 17.3%, or 18.3% after excluding one-time expenses related to the closure of a plant in North Carolina. Shaw Wu, an analyst at Kaufman Bros., said he was expecting an 18.6% profit margin in a Nov. 19 research note.

Dell’s consumer sales fell by 10% during the quarter, but Gladden said Dell “walked away from some retail business during the quarter” that wasn’t acceptably profitable in order to preserve margins.

Turn back to BusinessWeek.com later tonight for a full report on Dell’s third quarter, and a look at what’s ahead for the company.

Salesforce will begin selling the new software, called Chatter, next year at a price of $50 per user each month. The software works with Salesforce’s core customer management software to display “profiles” of employees and posts about projects they’re working on or customers they’ve visited. “I know more about these strangers on Facebook than I do about my own employees and what they’re working on,” Benioff said during a speech at the company’s Dreamforce conference in San Francisco. “I know when my friends went to the movies, but not when my VP of sales visited our top customer.”

Chatter pushes Salesforce, expected to reach $1.3 billion in revenues this year, into the crowded field for collaboration software. Salesforce is trying to expand beyond the customer management software that’s been its bread and butter. Microsoft’s SharePoint Server, an IBM product called Atlas that works with its Lotus e-mail software, and Google’s recently introduced Wave all offer business users the ability to share information and hold conversations on the Web.

Software developers will be able to use Chatter to build their own applications, Salesforce said. The move comes as some of the tech industry’s largest vendors are releasing tools that let programmers create cloud computing applications delivered over the Internet. Microsoft on Nov. 17 launched Windows Azure, software for letting Windows developers build cloud computing applications using familiar Microsoft technologies. Google and Amazon.com also offer tools for developers to build cloud applications.

Look for updated coverage on BusinessWeek.com, including excepts from an interview I’ll conduct with Benioff later today.

The unanimous "declaratory ruling" made good on a promise FCC Chairman Julius Genachowski made in an otherwise coolly received speech at an industry conference in early October. Under the new rules, state and local governments must act within 90 days of receiving an application for a co-location, that is, a tower site to be shared with other operators, and 150 days for other applications. Carriers have complained that governments are frustrating their efforts to improve coverage by sitting on tower applications indefinitely.

The FCC also ruled that state or local governments may not use the fact that wireless service is available from another carrier as ground for rejecting an application. And they may not require a zoning variance for every cell site.

For carriers, this additional fee spells a revenue bonanza. An average American pays $50 in wireless service fees today, according to industry association CTIA. As consumers tuck on additional data services, such as those for their new laptops, netbooks and smartphones, that amount could begin to climb, even if voice minute charges keep on shrinking. Average monthly bill amount has been essentially flat since 2003.

Money transfers' popularity shouldn't come as a surprise. This is a huge business in the U.S. and around the world already. Today, most consumers have to come into, say, a Western Union location or use their PC to complete a transaction. It can be problematic and time-consuming for many people to get to either one; but most everyone carries a phone. Another consultant, Juniper Research, expects international mobile money transfers to top $65 billion by 2014.

For the top 10 features Gartner expects to be present on smartphones in 2012, check out this list.

Factery Labs, debuting early Nov. 17, aims to pick up where Google leaves off. Instead of providing the usual list of Web pages, the Menlo Park (Calif.)-based startup reads all those pages first and then extracts facts from them by zeroing in on sentences--strings with a subject, then a verb--and assuming they represent facts of some kind. Then it creates an index of those facts and ranks them. The technology is called FactRank, in a nod to Google's patented PageRank.

"People want facts" out of their searches, says Factery Labs cofounder and President Paul Pedersen, a veteran of search engines such as Infoseek, Google, and Powerset and founder of data management firm Mark Logic. "They want to know right here and right now."

Factery is focused at least for now on providing search results for real-time and social sites such as Twitter, Facebook, and Yahoo's Delicious. Pedersen says traditional search engines don't work as well on real-time and social Web services because increasingly at least some people care more about what their friends and colleagues and other people they follow think is important news. Factery also isn't yet offering a full consumer search engine, instead forging partnerships with companies such as social network manager Sobees and encouraging other developers to use its technology.

Search for "Barack Obama" on Google and you get Google News links, Obama's Organizing for America site, a Wikipedia listing, and other links, but only with snippets of sentences. Search on Factery's FactFinder demo, and you'll get separate lists of facts based on Twitter recommendations and on Yahoo links (through Yahoo's Build your Own Search Service platform), such as "Barack Obama met students from Shanghai on a three-day visit to China," and "Barack Hussein Obama was sworn in as the 44th president of the United States on Jan. 20, 2009." While you still have to sift through those facts for whatever information you're looking for, the point is that you don't have to go through each Web page to find those facts.

Not all of these are provable facts, of course: One Twitter Recommendations "fact" mined from a site called hotnovels.info reads: "Barack Obama once roundhouse kicked somebody so hard that they went back in time and crashed into Amelia Earhart’s plane." Right. But there are interesting nuggets.

On mobile devices in particular, Pedersen believes, a search engine that provides facts is potentially much more useful than traditional search engines. It's still cumbersome to click on links using tiny keyboards, and then follow those links to find what you're really looking for.

There's one downside to Factery's system, at least for now, and it's a big one: At least on the demo site, all that scanning and analysis of pages takes awhile to produce results--several seconds, in fact. Clearly, Pedersen concedes, that's too long, and he promises he will reduce that latency. He won't get very far if he doesn't.

As for how Factery will make money, Pedersen says that a ways off. But he envisions something like "sponsored facts," whereby a General Motors, say, can pay to display a fact that is (in fact) an advertisement. And of course, it can do regular search ads like Google does.

Factery Labs received $1.2 million in funding in September from U.S. Venture Partners, Pedersen, and angel investors including super-angel Ron Conway. The cofounder is Sean Gaddis, who also worked at Google and Powerset (bought by Microsoft), as well as Yoomba, eBay and its Skype unit.

Given how many search upstarts have failed to gain any traction against Google, Factery Labs remains a long bet. But it's reassuring that the industry doesn't have to depend only on Google to push the envelope.

But the question is what happens to Droid sales in week two, three and those that follow--as dozens of other Android phones are expected to hit the market, as Olga Kharif points out in her story in the magazine. As of now, the device--which has gotten mostly good reviews--is benefiting from a huge marketing push from Verizon, estimated to total around $100 million. No doubt, you've seen the "I Don't" ads, which clearly position Droid as a superior alternative to Apple's iPhone.

That's an effective advertising campaign, especially when combined with Verizon's "there's a map for that" ads that bust on AT&T's reputation for spotty 3G coverage. Former Motorola CEO Ed Zander, for one, thinks Moto "has a good shot to sell a ton of Droids" if the device emerges as the gotta-have phone on the Verizon network. Indeed, if Motorola can maintain this 250,000-a-week clip for a quarter, it would move 3.25 million Droids. That would make it a blockbuster and the iPhone's nearest rival. Apple sold 7.4 million of its iPhone 3Gs in the company's just announced fiscal quarter. And Flurry's Peter Farago says the firm's data shows that Apple sold 600,000 iPhones during Droid's debut week.

But Droid's main competition isn't really the iPhone: it's fragmentation of the Android market. Clearly, Apple will have no problem keeping consumers focused on its device. The iPhone is the only smart phone Apple sells, and the company spends beaucoup bucks reinforcing a clear, powerful message: buy an iPhone, and get the benefit of Apple quality as well as those 100,000 apps in the App Store.

Now consider Motorola's challenge. Within weeks, consumers who go into a Verizon store will have many of different phones to choose from. Many of these devices will have a different "skin", a layer of software interface to make it stand out. That may make strategic sense on paper, but all these different interfaces is bound to confuse consumers. Also, it's not clear to me whether all of those 12,000-plus Android apps will run on all Android devices, further muddling the message.

And Zander wonders if consumers will be put off by the complexity of the Android model. It's bad enough with the iPhone, where Apple is responsible for the device and AT&T for the network. With Android, "are you buying from Verizon, or Google or Motorola?" While Zander thinks current Motorola CEO Sanjay Jha has done "a helluva job," he also thinks "there are a lot of competitors in this space. It's going to be an interesting Christmas."

In my lead story for the package, "Fertile Ground for Startups," I made two big points:
1. Startups are playing an increasingly important role in American business
2. Startups may play a central role in any recovery.

There was one startling new study, based on 2007 Census data, I was unable to work into the story that I want to highlight now, which provides some empirical evidence supporting the second point.

According to a new study by the Ewing Marion Kauffman Foundation, which was co-written by the respected economist Robert Litan, companies less than five years old generated nearly two-third of the net new jobs created in the U.S. in 2007. Without these startups, "net job creation for the American economy would be negative in all but a handful of years."

The upshot: It is clear more than ever that new companies and the entrepreneurs that lead them are the engines of job creation and economic recovery.

It is well known within economic circles that new companies produce the majority of new jobs in the U.S. economy. What this reports reveals for the first time is extent of that trend, and the fact that startups play a particularly important role in growing jobs out of a recession. New companies have produced all of the net new jobs in the U.S. from 2001-2007, and also from 1980-1983, the last big American downturn, according to the study.

This has huge implications for the Obama Administration as it tries to get the economy growing again. It also suggests a shift in focus and policy. Instead of focusing on saving large and dying industries (i.e. automakers and banks), the Obama Admin. should "begin paying more attention to removing roadblocks to entrepreneurs who will lead us out our current (well-founded) pessimism about jobs and sustain economic expansion over the long run," wrote the authors of the report.

One idea the authors float: granting a payroll tax holiday for new companies. This is a great and timely idea, especially since the expected health care reform law is supposed to increase the cost of doing business for new companies. Such a targeted tax cut could help offset the pain of health care reform and give new employers more of a reason to ramp up hiring.

This isn't to discount the role that big companies play in job creation. Large companies with 10,000 or more employees account for more than 10% of net job creation. But even then the authors note that those new net jobs may stem from the process of big companies acquiring young ones. "One of the only ways for big companies to add net jobs is to acquire younger companies that are not only generating jobs, but also are responsible for a good number of innovations that will keep the bigger company's revenue from diminishing," write the authors.

- Spencer Ante also publishes the Creative Capital blog. Click here to read more.

Where Will the Jobs Come From? Startups!

The original Bold 9000, which hit the market in May, 2008, has been the flagship of the BlackBerry line. The Bold 9700, available today from T-Mobile for $200 on a two-year contract and on Nov. 22 from AT&T, offers everything the Bold did--sometimes better and sometimes smaller.

Physically, the Bold strongly resembles assorted Curve models, though it retains the Bold's premium look and feel with a metal bezel around the top of the handset and a leatherette back. Like recent curves, it replaces the fidgety track ball with an optical sensor that tracks finger movement on a pad below the screen--a big improvement. It weighs about half an ounce (14 g) less than then original Bold.

The keyboard is a bit smaller than the Bold 9000, mainly because the handset is a quarter-inch (6 mm) narrower. I found the shrinkage did little damage to my typing once I got used to it--and the narrower keyboard offered a surprise benefit. Holding the Bold in my left hand, as I usually do, it was much easier to stretch my thumb across to reach the delete key when scrolling through the inbox.

The Bold shares the new BlackBerry 5.0 operating system with the Storm2. The browser is improved from the original Bold, though the smaller display makes it less satisfactory than the Storm. BlackBerry is working oin a new WebKit-based browser that it badly needs, since even this newest version is far inferior to what is offered on the iPhone, Adroid phones, and the Palm Pre.

Otherwise, what you get on the Bold 9700 is pretty much what you would expect from a top-of-the-line BlackBerry: Outstanding messaging, a very good phone, and OK everything else. Like other BlackBerrys, the Bold 9700 was seamless at switching between the phone network and Wi-Fi, and using UMA technology, it is able to make ordinary voice calls over a Wi-Fi network. This is especially important on the T-Mobile version, since T-Mo's U.S. coverage leaves a lot to be desired.

Network consideration aside, the choice between the Bold 9700 and the Verizon Storm2 presents a fundamental tradeoff: The Storm gives you a big display, but you have to sacrifice the physical keyboard. So far, at least, RIM has resisted slider designs and personally I have yet to find a slider that isn't awkward to hold and type on when the keyboard is open. Until someone finds a magic solution, this will be a fundamental compromise required in all smartphones.

I'd heard that an increase of 10% to 15% would likely get the deal done, so this improvement seems designed to accomplish two simultaneous goals: to put the acquisition over the top, without sending the message that Cisco will panic and radically pay up when shareholders of acquisition targets hold out for more. That's critical for a company as acquisitive as Cisco, which has done four large deals in just the last 45 days. At Cisco's shareholder meeting on Nov. 12, Cisco CEO John T. Chambers warned that "I'll walk" rather than overpay. "We're not going to pay a price that we don't think is good for shareholders."

One way or another, Chambers needed to get this deal done. He has said that video is his number one strategic priority, and video-conferencing in particular is a great opportunity for Cisco. Few, if any, forms of traffic chew up bandwidth and require more sophisticated routing and switching than videoconferencing--which needs to be not only high-res, but real-time.

And buying Tandberg was clearly the best way to accelerate his grand video plans. The company is not only the market leader in videoconferencing gear, but it's by far the hottest player in the market--not only with the mid-tier conference room systems that are the bulk of the industry, but also for high-end telepresence systems like the ones Cisco sells. Multiple industry sources I've spoken to say Tandberg routinely beats Cisco in deals for these systems, which create the illusion that you're actually sitting in the same room with other attendees, wherever they may be.

Also, Cisco needed to find a way to embrace open standards for its telepresence offerings. Currently, Cisco's systems only work with other Cisco systems, for the most part. That's unacceptable, for a company that built its Internet equipment empire by championing the most important open standard of them all--the Internet Protocol. Analysts say Tandberg is a leader not only in product innovation, but in making its gear inter-operate with other brands.

Here's a video of Chambers and Tandberg CEO Fredrik Halvorsen talking about the deal at the time.

Glyde, a startup created by former eBay Motors founder and chief Simon Rothman, is betting there's plenty of room left for newcomers that can make the whole buying and selling process much faster and easier. On Nov. 16, the Palo Alto-based company is debuting the site, which it promises will provide single-click purchasing for buyers (after the first one, during which you have to fill in your credit-card and shipping information) and 10-second item listing for sellers.

Rothman landed on the idea after leaving eBay in 2005 and "decompressing" in his native Ohio, where he discovered many people--real people in the middle of the country, outside Silicon Valley's reality distortion field--not only didn't sell or even buy online but had no desire to do so. Too much hassle, too much time, too little trust, he found.

Yet he estimates that some $300 billion worth of unused media products are collecting dust on people's shelves, or $3,000 per U.S. household. "E-commerce doesn't seem to work for real people," he says. At the same time, he adds, while other kinds of Web sites such as social networking have evolved, "e-commerce hasn't really changed materially."

Rothman and his team, which includes Chief Technology Officer Mark Wong-VanHaren, a onetime cofounder of Excite and former Charles River Ventures partner, essentially have rethought the entire online buying and selling process to painstakingly reduce the friction throughout. They've designed their own search engine, payment system, pricing algorithms, shipping system (sellers receive a branded envelope for shipping), and more.

Indeed, in an era when Web companies toss out beta sites in a matter of months, the three-year-old company is a bit of a throwback. It has operated in secret--not even a mention on TechCrunch!--even since Glyde built the first version more than a year ago. It has been testing it ever since with an increasing number of family, friends, and friends of friends.

Glyde, which has 15 employees, got $6 million in 2007 in a Series A funding round led by Charles River Ventures. "Every layer of what they've built is changing the game," says Charles River partner Bill Tai.

After checking it out, including buying an item (which won't be delivered until Nov. 20), it's clear that the Glyde team put great care into the experience. It's a clean, uncluttered design, and the experience is indeed fast and easy.

I can't speak for the selling process yet. But a demonstration indicated it was much faster, as little as a few seconds. On eBay, it can take upwards of a half-hour, though to give eBay its due, that's because it does offer a lot more options thanks to its much larger product base. (Its Half.com unit that deals in used media is much closer to what Glyde's doing.) Glyde has no plans to compete in collectibles because it's impossible to standardize the data and selling processes for them. Most media products have data, images, and pricing information available to be aggregated.

Glyde charges sellers a 10% item when an item sells, with no upfront fees. The seller also pays $1.25 for the mailing envelope. Sellers are expected to ship items within one day.

Another interesting wrinkle is that Glyde isn't depending on eBay-style feedback from buyers to assess the quality of sellers. It can track whether the seller actually did ship on time, for example, and its algorithms can favor faster shippers in product listings.

The company has filed for a number of patents on various of its features and processes. So while its buying and selling processes look easy, Glyde is attempting to stake claims on them to build up a competitive advantage.

Whether all that's enough to attract people who don't buy much online, and the many more who never sell anything online, is hard to predict. Amazon's success has proved that reducing friction in the buying process works, and eBay (and more recently Amazon as well) helped many thousands of people sell things online much more easily.

Even with considerable improvements in the buying and selling processes, Glyde has a big job ahead first convincing people it's better enough to give it a try. And if it does get some traction, it must scale up the business on many fronts that have tripped up others in the past, from trust issues to payment processing. Still, it's one of the most interesting attempts in years to attempt to take e-commerce to the next level.

So-called orphan works, millions of books for which copyright laws still apply but whose rights owner is unknown or cannot be located, will still be scanned and sold in an online registry. New revisions to the plan call for an independent trustee to collect revenues generated from orphan works for up to 10 years, or until the rights holders are found. After 10 years, that money will be donated towards the continued effort to seek out copyright owners.

In September, head of the US Copyright Office Marybeth Peters said Google's initial "opt-out" proposal to scan orphan works before attempting to find rights owners amounted to a throwing out of "fundamental copyright principles." Though the most recent revisions stipulate more rigorous steps for collecting and distributing money to authors and publishers, the proposed agreement is still opt-out, as Danny Sullivan pointed out in his blog Search Engine Land. Peters is still likely to object.

Other revisions which are likely to sway some critics include a new geographical limit to the the deal. Now, the Google Books Settlement applies only to U.S., Canada, U.K., and Australia. That will please the governments of France and Germany, who have objected to the plan.

In the next week, U.S. District Court Judge Denny Chin is expected to schedule a fairness hearing to hear arguments for and against the revised agreement. Expect to hear from groups like the Open Book Alliance, the coalition led by former Microsoft antitrust watchdog Gary Reback, which has already objected to the new agreement in a blog post.

The only real surprise here was the swiftness and thoroughness of Apple's victory. Judge Alsup basically ruled that the OS X End User License Agreement (EULA), which prohibits the installation of the software on non-Apple hardware, is legal and means exactly what it says. It is just the latest in a long string of ruling upholding EULAs, sometimes called shrinkwrap or click-wrap licenses.

A hearing on remedies is scheduled for Dec. 14. The order does not cover several other claims by Apple, including breach of contract and trademark infringement, but the ruling suggest that Apple would be heavily favored to win should the remaining case ever come to trial. There is also similar litigation pending in Florida, where Psystar is based.

If you want a Psystar Mac clone--there are six models starting at $600--you probably should order it soon. I don't think they are going to be around for much longer.