Media Dojo

Archive for June, 2009

Guilty as charged for light posting since the AWS gathering two weeks ago. Closing on the sale of my house long-distance has turned into a voyage into the surreal. Yesterday, I told one of the bankers of my intention to be one of the first in line for Tim Burton’s re-make of Alice in Wonderland, which will have Johnny Depp playing the Mad Hatter. I figured I owed myself a dose of reality after falling down the rabbit hole of US real estate circa 2009.

Thankfully, tomorrow I’ll be going to SF for Structure ‘09. Om Malik and the crew at GigaOm have put together the right people for anyone needing the best single, slurp of cloud computing. Naturally, some of the usual suspects like Werner Vogels, Marc Benioff, Russ Daniels, Lew Tucker and others will be there. But there’s also a good selection of “middle layer” companies who can fill in a lot of the white space.

Media oriented players will be thin on the ground. But we’re still in the stage where it’s more important to know what’s under the hood than it’s exact fit into the media and marketing worlds. They’re running at wildly different speeds but are converging (colliding?) rapidly. I’ll be there as Media Dojo to corral more interviews and participation in my summer-long project, which will be a Cloud Computing Guide for the Media and Marketing Industries. Trips off the tongue, huh? Right now, it’s more important to be clear than creative. I’ll inject chutzpah into the title as we get closer to publication, which will be near end of Sept./early Oct.

Expect a raft of posts based on that event and more information about the MD Guide over the next few days.

Compared to real estate, I’m glad to be back working with something tangible like cloud computing.

The last company to present during last week’s AWS event was Zumobi. Spun out of Microsoft Research in 2006, Zumobi is betting on superphones such as iPhone, Android and Palm Pre as platforms for mobile application-based advertising. According to John SanGiovanni, Zumobi co-founder and VP of Product Design, superphones sport full fidelity browsers, robust SDKs, 3G speeds, as well as GPS capability. Given superphones’ capacity for high-end processing and rendering, branded mobile applications rather than display banners constitute the most important mobile advertising inventory.

Zumobi got its first taste of cloud computing with AWS through the 2008 Summer Olympics. Lenovo and Intel were the main sponsors for branded mobile apps that needed to stay in synch with the evolving action in Beijing in terms of updating scores and provoding other context tual information. With the large amount of stored data that needed to be accessed rapidly, Zumobi started out with Amazon S3 as its primary data store for the Olympics. That experience led to the next branded project which was working on the Xbox 360 launch for Microsoft. For that project, Zumobi tried out EC2 for the first time. They also worked on their own back-end to tune some of their internal load balancing systems to accommodate the Amazon infrastructure. Then came work with American Idol which pushed the AWS partnership harder as scale and speed requirements co-mingled. By that time, roughly 7-9 months ago, Zumobi decided to port nearly all of its operational infrastructure over to AWS.

Aside from the evolution of the Zumobi/AWS relationship, John focused on how to parse some of the blizzard of iPhone and other superphone statistics spit out by the research industry and the media. There may well be over 1 billion apps downloaded from the iPhone app store. But the vast majority of these are “transient” apps (e.g. beer sloshing and fart sounds), meaning that they live isolated on the mobile device for a limited period of time and then are uninstalled. There is very little scope for network interaction.

Zumobi places its future on mobile applications in which there is a strong content anchor. By that, John meant that there is a recurring, refreshing dose of content that keeps the app active and conversing with a network service. Hence, a branded app from REI that allows a skier to source snow conditions on selected resorts fits the criteria. Zumobi partners heavily with media companies to get rich recurring content to fuel the app, drive the engagement with the user, and increase the value of the mobile app to a potential sponsor.

The issue brought up by superphones is that now, the bar is raised for recurring content to include video, real-time data, images all of which point to massive scalability issues with a mobile app. “In order for us to build a network of superphone applications, having a flexible data center is absolutely imperative”, he says. In concrete terms, John said that going wholesale with AWS has eliminated fixed costs (as a matter of course) , and save around 80% in variable costs. Along with the costs savings , SanGiovanni notes that Quality of Service, Quality Assurance, and geo-location functions have been greatly improved.

Of cource, whether this is a match made in heaven will be decided once Zumobi launches a major app right as AWS has a hiccup. I’ll be curious to see how the company plans for disaster recovery having put so many eggs into the same basket. That said, the fact that AWS was able to produce a clean example of a customer putting their operational destiny on the line is worth noting.

Time for the second installment of last week’s Amazon Web Services (AWS) meeting for local start-ups in Seattle. Rounding out the four customers presenting last Thursday were two Seattle-based media plays, nuTsie and Zumobi.

First up was Bob Wise, VP of Engineering of nuTsie, (www.nutsie.com), which allows people to port their iTunes play list across web and mobile platforms, including Blackberry and iPhone.  Basically, nuTsie takes a user’s existing iTunes library and rolls it into a streaming service much like Pandora. They don’t use the actual music in the library but the meta data about the songs and/or a playlist to create a super customized experience anytime, anywhere. If it seems a little disjointed there is method to the madness. Music licensing remains a mess even after a decade of industry tinkering. Like Pandora, Melodeo must make all its music streaming DCMA compliant so legally nuTsie is considering web radio rather than a a formal music distribution service.  The primary outlets are streaming for the web and for mobile phones. The business model is based on advertising for web streaming and subscriptions mobile phones.

For plumbing, Melodeo uses Amazon S3 to store and serve up the audio files (several TB in aggregate) that stream via a Flash player. The web-based nuTsie service gets about 10 million page views per month with about 10,000 hours of streaming music content served up each day between the web and mobile components. Both the streaming service and the mobile play are hosted on AWS. Bob said that for a typical load, it takes about 40 EC2 instances (think 40 virtual servers) that are about evenly split between large and small instances with one extra large instance for the main database. If you do a back of the envelope calculation it works out to roughly $10-15 per hour for pure compute capacity. Remember that nuTsie is also paying for data and certain transfer bandwidth charges.

source: http://aws.amazon.com/ec2/#pricing

One aspect of Bob’s presentation I liked was how he illustrated the effect of business forces on technical design. Chris Anderson of Wired fame used music as exhibit A of his Long Tail hypothesis. Bob said that in his experience the long tail might be long but it’s also thin as fishing line. Basically, this means that ultimately the number of music plays instead of the number of music tracks is what makes or breaks the business. Given the fact that the action stays with a relatively small number of tracks, nuTsie uses Amazon S3 as a content delivery network (CDN). If it sounds strange to use a data storage service to serve up content, take a look at charging. With many other CDNs in the market, a business is charged according to how much data sits at the edge node plus the transfer bandwidth to the end user. Thus, the key cost point is how much you get charged for keeping music tracks in storage which aren’t being played very much. Sticking several TB of music data out there on various edge nodes is an expensive way to do things. If you look at parking data similar to parking cars, loading rarely played music or video on an edge node is a bit like using a parking meter or a temporary lot whereas oft-played content needs the equivalent of a monthly reserved space. It’s an imperfect comparison I know. However, it’s decently clear that some of the heavy lifting for media providers is to figure how thick is the head of their demand model and how thin is the tail. Otherwise, it’s money out the door, cloud or not.

The Mariners were practicing today at Safeco Field where Amazon Web Services held a developer and start-up afternoon. Adam Selipsky and Matt Tavis of AWS covered the business layer as well as what was under the hood. Four customers gave their perspective about how they use AWS. Trains passed by as did the occasional crack of a bat to add background music to the presos.

I came away from the afternoon thinking that cloud computing has shifted into 3rd gear. We’re not yet at freeway speeds but it is clear that a lot of the theory is being hammered into practical application. Adam and Matt’s job is to solidify the current AWS story. Given that I’ve seen Werner Vogels recite the AWS lineage on numerous occasions, I can’t say that I heard anything new. Paradoxically, that makes me feel good. It suggests that AWS is in the midst of the tedious though critical task of truly productizing the offer. It’s easy for snarky analysts to harp about cloud computing being another form of time sharing, something we’ve had since the 1960s. But while it’s one thing to understand what something is, it’s quite a different challenge to make that something work at scale. Kind of like the physicist who can model a fastball. Very good, professor. Now, hit the sucker when it’s thrown at you at 96mph. Thus, I’m not particularly fussed when I hear AWS repeat its basic story with some incremental additions thrown in. Indeed, I’d say their success measure by 2015 is to become as boring and crucial as the power companies they seek to emulate.

To me, the customers provided the more exciting content. Jeff Lawson, Twilio’s CEO, spoke about how they offer telephony as a web service using multiple cloud providers, not just AWS. Strip out the black magic, and Twilio is about making telephony app development accessible to web developers. So if your expertise isn’t SIP, 3GPP or some other exotic telecommunications protocol, no worries. You can add various telecom services to your web app using traditional web standards and tools. Cloud infrastructure, much of which is AWS, enables Twilio to split its internal world into three large domains, continents if you will. There’s a DEV zone in which Twilio’s staff developers access a simple, powerful API that has only five blocks for building voice mail systems, IVR, PBX, click-to-call and other telecom services. Then there’s a STAGE zone that lets the developer test the app. Then there’s a PROD zone which is the only cluster that touches external customers (e.g. web developers). Media providers and marketers need to take note that branded comms is a huge future growth and community building area. Nearly every marketer I know bleats about how they’re fed up with using page views as a negotiating currency for media buys. They want more engagement (whatever the hell that is) from consumers. Nothing seems more engaging than a direct conversation, either between a brand and a consumer or consumers communicating in the context of a branded environment. I can’t say that Twilio solves the engagement problem (let alone how to use telephony as an ad currency). But you don’t get more guts level customer dialog than when it’s spoken instead of written so something is bound to shake out. Cloud now makes it far more likely that a savvy web developer will catch lightning in a bottle by using telephony in a new way. Watch this space.

However, you can bet that serving up telephony or other latency-sensitive applications beg for more robust testing. John or Jane Q Public won’t be thinking, “oh well, the app’s still in BETA. They’ll get it right with the next rev.” Far more likely, ordinary people will think “this sucks!” and move on when they encounter a communications or media app that is late in delivering the goods. Enter SOASTA, which uses the cloud to be an Underwriters Laboratory equivalent for websites and web based apps. They remind me of those demo guys who bring down city office buildings with explosives except SOASTA’s job is to stress test major web sites for latency issues. A case in point is Intuit. Many remember April 15, 2007 when Turbotax melted at 1015pm, not the company’s best day. The problem, however, is that to get a proper test, you need to simulate a proper load. SOASTA uses AWS and other providers to simulate massive traffic loads (eg. 300,000 simultaneous users) in various combinations without the need to construct a separate test facility. According to SOASTA’s CEO Tom Lunibos, the dirty little secret of most Web 2.0 apps (1.0 for that matter) is that they were lit without having done more than cursory load testing. The idea was that you got it out there, had it melt, apologized to the users with some cute stick figure with a hard hat, and fixed your latency problem. Mainstream adoption and recessions are curing that stunt pronto. Tom  declared (rightly) that latency can be measured as lost sales. Outages in 1998 that were reported in PC Magazine of Information Week now grace the front pages of the New York Times and Wall Street Journal. Bottom line for media providers (especially gaming companies) is that latency is money. So blow up your pre-launch site. You’ll be glad you did.

Don’t want to get too long so I’ll follow up with Melodeo (cloud-based music serving) and Zumobi (mobile applications as marketing) in a separate post.

It’s hard to discuss Google without having a general discussion about Internet Media/Marketing and vice-versa. I’m not sure the same dichotomy exists in the cloud world yet. But it’s safe to say that you can’t have a comprehensive cloud computing policy discussion without referring to Google. Hat tip to the Knight Center for Specialized Journalism for bringing Google’s regulatory policy lead to speak to us about the web giant’s role in policy making.

Alan Davidson is the Government Affairs point person for Google in Washington. Naturally, he’s a computer scientist and a lawyer by training.

Alan believes that we’re currently in the Empire Strikes Back part II of the policy trilogy. The first part of Internet policy had a certain naivete on the policy side, summed up by the thinking that the Internet was a particular “thing”. It was a new technology system to be sure, but one that could eventually be wrestled into an existing policy tool kit. Step change is a nice theory but rarely observed in practice. At the same time, digital utopians believed that all the new users adhered to similar norms that marked the early generation of “netizens”. Given the information asymmetry in favor of the technologists and early adopters, it shouldn’t be surprising that advantage went to the rebels.

However, the last few years have seen regulatory authorities world-wide respond to all the John Perry Barlowesque declarations of cyber-independence with, “You don’t think we can regulate the Internet? Well, watch us…” In the past two years, major Google services in 23 countries have been blocked outright. For example, YouTube doesn’t exist in Turkey ever since a Greek activist posted a parody video claiming that Turkish independence icon Attaturk was gay. The Turkish Government told its ISPs to block all YouTube traffic into Turkey full stop. Similar incidents with various Google services have occurred in Pakistan, China and Bangladesh.

Alan was quick to caution that regulation itself isn’t inherently bad, however. Think for a moment as a U.S. citizen how the 4th Amendment regarding unreasonable search and seizure applies in a cloud-based environment. If you keep your correspondence and/or your business records in your house or on your person, the police or other authorities have numerous procedural hoops to cross before they can seize them. Not so once you port all that stuff into a cloud. People do it all the time with Quicken Finance, which is convenient to be sure, but buried in the privacy policy is the well-worn phrase that the company will protect your information up to the point it receives a “lawful request” from the “proper authorities”.

Knotty policy question often expose the cultural rifts between Silicon Valley and Washington. Alan freely admits that since Google is an engineering company, its first response to almost any problem is to try to figure if there is a technical means to solve it. Washington looks for a law. The Valley builds a tool. Occasionally, the twain need to meet.

The session ended with me asking the obvious question about cloud computing policy—taxation. Privacy is an issue that will require decades of piling mess upon mess until something workable emerges. Taxation of cloud-based commerce is a hear and now that hits all pockets. States are desperate for revenue in a recession. Many are scrambling to introduce Internet commerce sales tax. The issue boils down to a simple question of nexus: where did the “commerce” take place? Did it happen in Google’s data centers on the Columbia river? What about a small publisher in North Dakota using App Engine, Azure or EC2 that’s pulling data from S3 or another data provider in another state? Then you have a customer in New York who’s ordering and accessing the content through their New Jersey-based ISP. The bottom line is that under classic conceptions of nexus, upon which most sales tax theory is based, you could end up with 6-8 states having a claim on the potential sales tax revenue. Amazon and New York State are currently duking it out over that issue as it applies to Amazon’s current business.

Many, if not most, of these issues were argued when Internet commerce could be restricted to specific servers in specific states doing specific things. However, once you start looking into public cloud services that are selling both digital services and digital goods, you fall down the rabbit hole. The smart money right now is talking about a Federal Excise Tax on Internet commerce, which will hit cloud services. But what one hand simplifies (e.g. single rate, single collection point, single collector) the other hand flails about ( does a call to a S3 account in one state count less than spinning an EC2 instance in another). The only sure bet is that if cloud computing reaches its lofty goals for adoption, the bars in D.C. will be packed with lawyers and activists talking cloud policy for some time to come.

I’ll post some more about the D.C. trip with the Knight Foundation after working through some more notes. In the meantime, here’s a shout-out for the upcoming Structure ‘09 conference being put on June 25 in SF by Om Malik’s GigaOm crew. Regardless of the policy hairball that will soon grip the industry, it’s mega important to get all the top players in one room (can still do it for cloud computing infrastructure). That’s what they’ve pulled off. I’ll be there cranking out copy that affects the media and marketing industries.

http://events.gigaom.com/structure/09/