Author: Steve Wildstrom

Netflix has become a big star by moving from its successful business of lending DVDs to streaming to the even greater success of its own productions. What’s less known is that Netflix has become a technological leader, teaching the rest of the industry how to run a vast system in the cloud that has made it a winner.

Last year, the company had revenues of $5.5 billion, up 25.8% from 2013. Profits were $267 million. Streaming delivered 88.9% of the business though it has only been in 2007 that Netflix’s business consisted entirely of shipping DVDs and Blu-rays.

But another place Netflix has by itself is as a huge and extremely clever, successful user of Amazon Web Services (AWS), to which it moved its operations in 2010. Any Netflix service a customer gets — from signing up to a subscription to making the service available on just about any device capable of displaying the content to watching a showing of “Orange Is the New Black” to data and streaming is handled on AWS.

Big distribution, big failures. A massive distribution system built on the cloud seems asking for big failure. In 2011, AWS had a large breakdown that cut off the services from a good number of cloud participants including Reddit, Quora, and FourSquare. Netflix kept its service intact with almost no loss. And it wasn’t luck; Netflix has gone on to suffer virtually no losses during subsequent network problems.

I’m not sure how many users with cloud bases services (even private clouds) use the Netflix-developed techniques. Netflix is so pleased with the processes, it is happy to share what it has learned to anyone who wants the information. Not only that, the critical software of its diagnostic and control service is available freely through GitHub.

From the beginning, Netflix laid out its intention of using AWS. John Ciancutti, then a Netflix vice president and now chief product officer at Coursera, wrote a blog post (( If you are at all interested in the Netflix effort, follow the blog. It contains a wealth of information on how to run a big cloud-based service. )) on why AWS was chosen. The four top reasons: flexibility in architecture of a growing service, letting Netflix engineers focus on the business while Amazon provided the data center, the ability to easily adjust the size of the system even if the company itself isn’t good at predicting, and a commitment to the cloud as the future. [pullquote]The design of Netflix operations on AWS and the means for testing them was in large measure because of the warning from Amazon CTO Werner Vogels: “Everything fails. All the time.”[/pullquote]

The Simian Army. The design of Netflix operations on AWS and the means for testing them was in large measure because of the warning from Amazon CTO Werner Vogels: “Everything fails. All the time.” A major accomplishment for Netflix has been a large package of routines with the quaint name of the Simian Army. It was getting organized just around the time of the 2011 AWS network failure and has been expanding steadily in recent years with additional functions such as security tests. It’s quite notable that the inventions of Netflix are recommended to its cloud-using customers by IBM.

The oldest and one of the best known is Chaos Monkey, a tool available since 2012 to anyone who wants to use it. It is designed to improve the system by injecting failure, teaching engineers how to make the network service the daily threats with minimal loss.

“The name comes from the idea of unleashing a wild monkey with a weapon in your data center (or cloud region) to randomly shoot down instances and chew through cables — all the while we continue serving our customers without interruption,” Yury Izrailevsky Ariel Tseitlin wrote in the blog. “By running Chaos Monkey in the middle of a business day, in a carefully monitored environment with engineers standing by to address any problems, we can still learn the lessons about the weaknesses of our system, and build automatic recovery mechanisms to deal with them. So next time an instance fails at 3 am on a Sunday, we won’t even notice.”

I’m convinced that Simian Army is one of the reasons behind Netflix’s excellence (of course, there is also the willingness to spend a lot of money on content and internet capacity). For whatever reason, there’s no indication Amazon itself uses Simian Army, though that may be just part of Amazon’s secrecy. In any event, I regularly watch both Netflix and Amazon Prime and the quality of Netflix streaming is better. It may be the quality of the network testing.

We’ve been hearing for years about how the internet would keep track of the refrigerator and let us know when we need milk and eggs. Now it is finally coming, as should surprise no one, from Amazon. The question is, will imposition of what we’ll call the Internet of Shopping be worth it to anyone?

Amazon started this approach  with the Amazon Dash, a camera and microphone-equipped device that used a WiFi connection to order goods from AmazonFresh, the grocery store-by-truck being offered in selected cities. But the Amazon Dash Dot, for which an Amazon Prime consumer can submit an advance order, is both a more complex and simpler service.

The button is linked to a specific product, at the moment at least from a collection of massive consumer companies such as General Foods, Kraft, and Procter & Gamble. Stick the 2″ by 1/2″ Button to a surface. When you are running short on Tide Pod detergent, Clorox Disinfecting Wipes, or Kraft Easy Mac Cups, the Button sends a WiFi signal to an iPhone, Android phone or a PC to generate an order to Amazon.

Once the Amazon Internet of Shopping device starts shipping in volume it is likely to do well. The Dash button will be easily available to anyone–it can be had by any purchaser for no cost and is likely to become popular.

Assuming that the button works as expected, and that seems like a realistic expectation, its real significant will be its economic potential. Amazon, starting with taking over the book business, has had a significant impact on a considerable variety of products.

In our house, we order lots of stuff from Amazon, at least enough that we get a delivery nearly every day. We’ve essentially replaced drug store purchases with Amazon, in large part because our neighborhood CVS is a pain of a store. But we–and by this I mean my wife, Susan, who does most of the ordinary shopping–still buy most basic home products in supermarkets. We want to buy our own fresh food (AmazonFresh is not an option here yet, though PeaPod or Safeway deliver) because we want to check out quality, availability and price. I personally prefer hitting the H Mart Korean store for produce and seafood.

The Dash button could make the Internet of Shopping a winner for both Amazon and consumer product makers. Amazon, of course, would increase its sales if the button succeeds. And product makers lock customers into the brand for which they have a button; it’s likely to get an order for needed detergent with Tide, dissuading customers from taking advantage of an in-store special. Products with firm customer loyalty for the brand, such as toiletries and cosmetics, are likely to do best.

I suspect the Internet of Shopping is likely to do best among relatively well-off families. Of course, the option only works when you have a WiFi connection at home, but that comes close to 100% when you eliminate the poorest and the oldest. But those relatively high in income tend to be longer on money, shorter on time and more interested in the ease generated by the Dash button than in saving.

The New Yorker offers a somewhat tongue-in-cheek view of the disadvantages of Dash button, “The Horror of Amazon’s New Dash Button” by Ian Crouch:

But what if there is actual value in running out of things? The sinking feeling that comes as you yank a garbage bag out of the box and meet no resistance from further reinforcements is also an opportunity to ask yourself all kinds of questions, from “Do I want to continue using this brand of bag?” to “Why in the hell am I producing so much trash?” The act of shopping—of leaving the house and going to a store, or, at the very least, of one-click ordering on the Amazon Web site—is a check against the inertia of consumption, not only in personal economic terms but in ethical ones as well. It is the chance to make a decision, a choice—even if that choice is simply to continue consuming. Look, we’re all going to keep using toothpaste, and the smarter consumer is the person who has a ten-pack of tubes from Costco in the closet. But shopping should make you feel bad, if only for a second. Pressing a little plastic button is too much fun.

That, of course, is a problem that only applies to the relatively comfortable. My guess is they are going to find Internet in Shopping too much of a helpful offer.

Mark Cuban is one of the inventors of internet-based video, having gotten rich selling Broadcast.com to Yahoo in 1999. He still admits he doesn’t know that it is desirable for all the new means of network distribution to dominate entertainment. He owns a television network, movie theaters, the Dallas Mavericks (where he bars fans from using phones or tablets while at an NBA game), and, of course, “Shark Tank” on ABC.

Why does he make such investments considering the background? Simple. He said in an appearance at the Code/Media conference (( Code/Media in Laguna Niguel, CA, was set up by the Re/code service. The conference was organized by Re/code’s media reporter and editor, Peter Kafka. )) on Feb. 18: “I don’t want my 16-year-old daughter to watch Netflix with her boyfriend. And I want to go to a movie so I don’t have to talk with my wife for two hours.”

Cuban was joking a bit, but there’s no doubt his judgment of business doesn’t matter. The folks in various entertainment/technology/internet enterprises tend to be committed to whatever line they are in. The rest of us are still figuring out what works.

Roku, Apple TV, and more. This jumble of services is delivered to smart phones, tablets, and PCs. In lots of homes, it comes to your TV through Roku, Apple TV, sticks from Google and Amazon, or sometimes features of a “smart TV” itself. For most homes, the internet device sits next to a cable box, but it is become an increasingly feasible choice to drop cable service and rely on the internet.

Kevin Tsujihara, CEO of Warner Bros. Entertainment, is one of the pushers behind industry UltraViolet, a program for cloud-based sharing of recorded Bluray recordings among multiple devices. But Disney, one of the industries most successful players, has developed its own, more successful solution. And the public has shown little interest in figuring our how to get the disks they have bought registered and replayed on UltraViolet. But Tsujihara still sees UltraViolet developing into a solution.

Internet-based entertainment appears to be based on two rival areas of attack. The set of mostly short-form video, the work of both amateur and professional programmers, appearing mostly on YouTube and Facebook and traditional video, mostly movies (including a lot of offerings that used to be know as direct-to-disk films) and TV shows, are being sent out through a variety of distribution, sometimes subscription, sometimes rental.

Chernin and Sling. DISH Network’s Sling TV is an extremely interesting startup because its program, in addition to a lot of competition with Netflix and other entertainment video, includes live sports events on ESPN and ESPN2—programming that, until now, was available only with cable service. Peter Chernin, the former president of News Corp. and now a production backer with the Chernin Group, told the Code/Media conference that he sees the Sling approach—likely to be offered by others—that live TV over the internet is the future.

“I’m not sure what the timeline is,” said Chernin. “But there’s clearly more money to be made online. Fifteen years ago people said sports would never be on cable.” (In a bit of a slip, he means internet, not the traditional TV cable channels.) And it is world business. In the U.S. we tend to think of football, baseball, and basketball. In the rest of the world, there is soccer and, as demonstrated recently, cricket, especially in India and Pakistan.

The existing TV networks are also pushing hard into the internet business as an outlet for current broadcast, not just recent shows. HBO has long made its programming available through HBO Go over the internet only as an alternative for cable subscribers. A new service is a direct subscription for those who want only internet service. Now NBC will soon offer the content of its owner-operated stations to consumers who want to subscribe only over the net in nine large cities, including New York, Los Angeles, and Chicago.

Internet regulation. One issue that remains a threat to how content is delivered over the internet has been the failure of regulation to keep up with the progress. So far it is working in a mostly acceptable way while politicians continue to argue over the future. But any day we could start to see issues where, for example, a network carrier such as Comcast or Verizon agrees to provide quality delivery to one competitor—or maybe just its own service.

The effort to settle it has sank into a political mess. The Democrats want internet carriers to be regulated like utilities. Republicans want pretty much everything left alone. Federal Communications Commission Chairman Tom Wheeler, a Democrat, tried to find middle ground but was jumped on by Republicans in Congress and his fellow Democrats, led by President Obama. Congress is now working on a regulation plan that probably will pass the FCC on a party line vote and be attacked in both Congress and the courts.

Cuban, as usual, took the tough role. He seems to respect Wheeler, but doesn’t think much of his effort to come up with a version of the Democratic plan because he will be replaced in a couple years by a new Democrat or Republican with unknown intentions. The action of the FCC “will f*** everything up,” he told the conference in his usual blunt way. “Having [political appointees] overseeing the internet scares the s*** out of me.”

Cyanogen Inc. is hardly one of the best known companies in the mobile phone business. It is a recent outgrowth from CyanogenMod, a not much better known developer of Google-free versions of Android for mobile devices. And now it’s making a move to turn the Android Open Source Project (AOSP) into a challenge to both Google and Apple.

There have been three interests controlling most of the phone world–Apple, Google Android, and AOSP. The Google version, which includes a broad range of Google apps, is dominated by leading Android phone makers, particularly Samsung. The AOSP approach is used on millions of phones sold in Asia, especially China, by Xiaomi and many others (it is also popular with a minority of Google device buyers who want to install their own software).

Tough moves. Cyanogen is now planning a tough move into the market. Microsoft has shown itself to be a investor in the project spending perhaps $70 million, though it has not made a formal commitment. Amazon is another likely contributor. (Ina Fried has written an excellent report for Re/code.)

Amazon’s potential is obvious. It has been doing its own AOSP version for the Kindle Fire tablets and the unfortunate Fire Phone. It could almost certainly benefit by managing Android without leaning on Google. The Microsoft game is somewhat more dubious. One possibility is it is just looking for an OS other than Windows or Google Android for lower-end Nokia phones intended for third world countries.

The other is to play to the advantage of Windows Phone, which has struggled as a distant competitor for both Apple and Google. The logic, which is purely speculative about Microsoft’s intent, is that injecting a third, well-funded developer into the Google/iPhone battle could help create a better position for Windows Phone.

Unlikely problems. The  Cyanogen move is unlikely to cause problems for Apple. The iPhone is designed only for the upper part of the market–remember the lack of success of the iPhone 5c–while whatever Cyanogen offers is likely to be aimed primarily at the low end. The addition of a new OS is unlikely to have much impact on Windows Phones either.

The big issue is the entry of a new layer into the market of Google’s Android. The relationship of Android between Google and AOSP is a complicated affair. (( For more on the complications of Android, see Ben Bajarin’s The Android Schism. ))  At first, AOSP was attractive primarily to tinkerers who wanted to tamper with their phone software. They replaced the manufacturer-supplied software with CyanogenMod versions, which were offered for a wide variety of Android phones.

But producers of phones in the rapidly growing Chinese market were strongly attached to AOSP. They wanted to get a free OS and they had little interest in installing the Google services, many of which would not work on the Chinese network anyway. ABI’s survey found AOSP software was responsible for about 20% of the smartphones sold worldwide this summer.

Cyanogen could bring  some valuable stability to the chaotic assortment of software and OS versions being sold. But there is a question about just what Google does with AOSP while a new rival company does its best t0 take charge of the system. Google typically has made each new version of Android available to AOSP (Honeycomb, the short-lived tablet version was an exception).

Google’s interests. With Cyanogen trying to take charge, Google my be far less interested in helping. Google’s mostly unprofitable Android business has been to gain use of the services built in to the phones. But Google is already being hurt by the growth of the disorganized AOSP phone network and it could suffer significantly more if Cyanogen turns this into a real, competitive business. Android software is available free for any would-be user, but Google is not required to maintain it.

Of course, there are still a lot of questions about how Cyanogen will make this a real success. Right now, its only business is to provide software for Micromax, a maker of low end phones in India. It can provide a business of helping manufacturers get a more efficient operating system from AOSP, but it is hard to see how that can produce a whole lot of business.

But it certainly can contribute to the already growing fragmentation of the Android world. Google’s interest in Android has tended to cycle, hitting a peak with the acquisition of Motorola and at something of a low right now. Android’s market continues to grow significantly, but the biggest part of the expansion remains on the AOSP side while the top Google Android makers, led by Samsung, are hurting. Google should be worried.

DISH Network is a company that has not gotten a lot of attention in the internet business. After all, it mainly exists as a cable TV competitor, suppling service to about 14 million customers through satellite dishes. But, pushed along by founder Charlie Ergen (left), it is moving in on the business, starting up new internet television services and elbowing its way into U.S. wireless channels.

The satellite-based internet service doesn’t draw a lot of envy from customers except those who have no better way of reaching the net. It has about 550,000 customers of what it calls “broadband” service, not enough to qualify for the term under the Federal Communication Commission’s new definition of 25 megabits per second. But DISH is making its way based on facilities of landline and terrestrial wireless services.

Aggressive DISH. The latest has been an aggressive movement into the most recent auction of wireless service. DISH is spending about $10 billion for wireless spectrum, most recently $3 billion in the FCC’s Advanced Wireless Services (AWS-3) auction. “That spectrum, combined with the wide-ranging spectrum licenses that DISH already owns, will turn Charlie Ergen’s company into a spectrum powerhouse,” writes Phil Goldstein of FierceWireless.

DISH is no stranger to controversy and this deal is no exception. It was able to get bargain prices on some of the AWS bids by working through two small companies, Northstar Wireless and SNR Wireless, which were eligible for a discount program. DISH owns 85% of the companies. According to the Wall Street Journal, Republican FCC Commissioner Ajit Pai will ask Chairman Tom Wheeler to investigate the action, which he claims “makes a mockery” of the small business program.

DISH has not yet laid out plans for the new spectrum it is acquiring, but it seems a safe bet it is not about to get into the wireless telephone business. Its aggressive bidding increased prices in the auction for competitors Verizon and AT&T. Maybe it’s hunting for a deal with the hungry Sprint or T-Mobile, or perhaps outright acquisition of one of these second tier companies. Or it could be building a network for a new land-based video service.

DISH’s new ideas. One description of the video idea is DISH has been looking for ways to expand its current TV services over the internet. In 2007, EchoStar, DISH’s parent company, acquired Sling (original model below), a device that accesses your cable TV system so you can watch service on any internet capable device in any location. ((Sling was the invention of Blake and Jason Kirkorian and sold for $380 million. One of its key functions was its ability, in pre-Netflix days, to adjust its service to capture the best available bandwidth.))

Although Sling remains a separate product, DISH has incorporated the technology in several ways. Hopper, originally a feature for moving cable service to different TVs in the house, now has Sling capability that lets you move content from your cable device to any device on the wired or wireless internet (as the ad points out, so the kid can watch the big game on his phone while making a visit to his TV-less aunt).

Now, a new service called Sling TV will be launched to compete head-on with an internet-based service available on PCs, mobile devices, and several plug-into-your-TV systems (detail for Insiders). A $20 monthly subscription will provide a variety of programing. The basic service is very close to what is available on Roku, Google TV, Xbox, and other internet-based services, either free in the basic subscription or, such as Netflix, at extra subscription costs.

But the difference that makes it a standout from the crowded field of internet delivery competition is the availability of ESPN and ESPN2. The inability of non-cable carriers to present live sports events has been a big disadvantage. Customer insistence on live sports kept them as cable service customers. While DISH has yet to demonstrate the success of its service, it has a real chance to make a dent in the existing cable business.

Cable competition. DISH, and Ergen, are well positioned for a cable and TV business that is changing in fundamental ways. The cable carriers are holding their share of the market–there have been slow decline in TV subscribers, but each year their landline business is developing more and more into providing internet (broadly defined as all TCP/IP traffic) rather than video channels. The wireless carriers are seeing less and less voice traffic and more and more data and video. Their growing LTE service runs on TCP/IP that is more like the traditional wireline but available anywhere. Cable carriers programming began mainly with broadcast channels and ESPN. Then it added enhanced networks, first paid channels such as HBO, then original programming on ad-supported ones such as AMC and Lifetime.

The companies that own most of the businesses–from CBS to Comcast to Verizon–move slowly, seeking to replaced their expensive services. It’s hardly a surprise HBO is prepared to offer its shows over the internet to buyers who are not already cable subscribers after Time Warner unloaded Time Warner Cable from the business. In a field marked by this turmoil and with real room for change, it is potentially a fine market for a daring company like DISH.

Microsoft’s latest in a series of Windows 10 show-offs yielded, seemingly as planned, attention for two display products under development — the 3D personal HoloLens and the giant Surface Hub. But the risk is not in these designs which, though very sexy, would cause little damage if they never made it to market. The challenge is Microsoft’s latest effort to unify its software as Windows across its product lines in reality, not just branding.

Microsoft has always been a bit ambiguous about the meaning of Windows. In the beginning, Windows was an operating system for PCs built on x86 processors (there were occasional stabs at other chips, even the PowerPC, but none survived for long). But versions varied. From the introduction of Windows NT in 1993 until the unifying release of Windows XP in 2001, Microsoft offered two families of Windows for PCs based on very different code

Early mobile Windows. The Pocket PC, a competitor of the Palm (( Microsoft wanted to call it the PalmPC, but was beaten down by the threat of lawsuits)), bore a resemblance to Windows in appearance but almost no relationship to the code. As Microsoft looked for the success in Windows Phone, the company fiddled with making it a winner. But it never got the display to work satisfactorily, nor found a way to unify the PC and phone code. Windows has moved closer with Windows 7 and Windows 8 and will get there with Windows 10 (9 has been skipped).

By itself, the change will be of little importance to Windows Phone owners. Even the developers of Windows Phone applications probably won’t see much change. The impact is on where the market goes from here. In terms of phones, Microsoft is starting with a new version of Office–Word, PowerPoint, Excel, Outlook–where a single code base serves PCs, phones, and any tablets that come along, and with UIs appropriate to the devices. In terms of user functions, I suspect the difference will not be very great and I still expect the function of, say, Excel, will remain limited to the small display of a phone.[pullquote]Windows is a bit odd from other approaches to operating systems[/pullquote]

Windows is a bit odd from other approaches to operating systems, however. Almost everything except BlackBird’s software is built on Linux or, in the case of OS X and iOS, Linux’s cousin BSD Unix. But Linux is just an operating system in the classical sense; the file systems, user interfaces, and everything else are done separately. Android is built on Linux.

Unix’s children. OS X and iOS are both built on Unix, but are separate code from each other. They share some APIs but nowhere near all of them; for example, OS X does not have the phone functions and screen touch that are critical to iOS. But Windows will have the same full OS for all products.

Although Microsoft has said this, it is not completely clear what it meant. Microsoft Vice President Joe Belfiore tweeted information explaining the difference between “desktop” versions of Windows for PCs and any devices with displays bigger than 8″ and the “No dsktp” version for phones and smaller devices.

The veteran Windows watcher Mary Jo Foley of ZDNet did her best to make sense of what she calls “the muddy waters.” One problem is Microsoft is coming up with so many APIs for different products. There is one basic version designed for devices running on ARM processors, including Lumias and other manufacturers’ Windows Phones and any tablets with displays less than than 8″. What Microsoft refers to as the Desktop SKU ((The use of SKU, which stands for stock keeping unit, is mysterious. It is normally a retailing and wholesaling term used to describe specific versions for stocking and pricing, but Microsoft seems to use it for versions of Windows used for different products.)) are devices including desktops and laptops as well as larger tablets. It’s not clear how many versions there will be for consumers and business customers, but there will definitely be special versions for HoloLens and the Surface Hub. (Microsoft has not yet talked about Windows Server and its code relationship to Windows 10.)

The fate of Surface RT. There will also be a Mobile SKU, described by Foley: “The Mobile SKU is aimed at devices with small RAM and disk requirements. It’s built for locked down devices, though, in theory at least, it could run on a device with any size screen.” One question is whether Microsoft will find a way to revive something replacing Surface RT, a sort-of tablet running a stripped down Windows 8, without the Win32 component, which permits the use of pre-Windows 8 versions.

Windows RT was a flop, both on the Surface and a handful of other tablets. One of the major flaws was the fact that pre-Windows 8 applications would not run and new Metro apps did not develop quickly enough–or even by now–to make a Windows version that couldn’t run them be attractive. Would the version of Windows 10 for the 10″+ tablet run Win32? Microsoft’s division at 8″ also rules out the approach of Apple, along with Samsung, HP, Lenovo, and other makers of Android tablets, of offering functionally equivalent tablets with screens between 7″ and 10″.

This all gives the evidence of why the operating systems story at Apple is so much simpler. iOS goes on the iPhone, iPad, and what’s left of the iPod (a stripped-down iPhone). The Apple Watch will use a version called, of course, Watch OS. One version of OS X runs on all Macs. Apple got out of server hardware and Xserve in 2010 and abandoned the Mac Pro Server in 2013. It now offers the OS Server, which is just a $20 add-on of server management for Yosemite intended for using a standard Mac for low demand server applications.

It would be nice if Microsoft’s offerings were as simple as Apple’s but the world of Windows is much more complex and, in the line of PCs, much bigger. We haven’t even seen the pricing plans yet (except for the free upgrade of Windows 8) but we can expect it will be expensive. But Microsoft doesn’t have the beautiful, simple freedom of supplying an OS to its own hardware and always updating it without charging users.