Updated: 1/6/2004; 11:10:47 PM.
Jeremy Allaire's Radio
An exploration of media, communications and applications over the Internet.

This is a personal weblog. The opinions expressed here represent my own and not those of my employer.


Thursday, April 24, 2003

Jeff Huber, VP of Tech for eBay

eBay generates $681 per second.  How does ebay compare to other commerce players?  They're now 31st largest commerce provider, bigger than GAP, ToysRUs, and others.

Where do they focus?  Front end (e.g. new and scarce products), end of life, and used/vintage products.  Asser that this is nearly a $2 trillion market opportunity.  Continue to see strong growth in most categories, but big in cars, home and garden, clothing and accessories.  Operate 27 sites globally.  Strong international growth.

<jeremy>Not sure what the relationship of this is to broadband infrastructure?</jeremy>

They're seing 6.5 million web service API calls per day!  Sending 21 million emails a day, conducting 80 million searches a day.  30% of inbound products are coming in through virtual supply chain --- e.g. web services.

With regards to broadband, it's the always-on nature that drives the nature and velocity of usage of eBay, and the ability to have an immediate, notification driven relationship with buyers and sellers.  From a bandwidth perspective, they can really improve the richness of the product views and details with broadband focused content.  And broadband empowers more individuals to create eBay-based businesses.

They'd like to become an alternative distribution channel for any manufacturer or retailer, and using multiple pricing models, including fixed price increasing dramatically.

Question:  web services adoption; how's it growing/going?  what standards are you using and are you ahead of the standards?

We haven't been as high-profile about PR, but activities are ahead of expectations in terms of adoption.  Went for a least common denominator approach --- e.g. it's not SOAP, but XML/HTTP and a RESTful API.

Question:  what impact on broadband content will happen?

I think one concept that came early today is the idea of a tipping poing, but they're still only at 50% of their end-users having broadband, so sellers are reluctant to use broadband rich media content.  Typically, the more expensive and advanced product areas are using broadband ready content.

Question:  how has the reputation system on eBay supported commerce?

One of the key issues they see is customer safety around the experience.  People just don't trust the system in general. Trust and safety systems, including reputation, are a big part of that.  Fraud management systems are a part of that, too.

David Labuda -- CEO Portal Software. 

Broadband will surprise us.  The apps that take off were never predicted.  Examples --- eBay and commerce; SMS and Mobile; ring-tones and mobile.  Nobody knows what the killer apps will be.  Must be able to launch services quickly and inventively, and fast experimentation.

The ability to intuitively price these new premium services is a real challenge, especially for communications companies used to just charging for data rates and time.  There really isn't a mapping b/w bandwidth used and value of the service.  This is a real sea change for a lot of these companies.  This is value-based pricing.  This could enable QoS models and cost structure savings to increase margin on high-value items.

Security and fraud prevention is a major issue still.  This is paramount because otherwise the communications companies will have terrible experiences and will back away.

Question: In favor of value-based pricing, but some of the pricing will be above and below the cost of bits.

Lots of people will try and beat the system, and you can try and reduce that problem, but it won't be the norm.  Agree there will be a variety of low and high value transactions, but that for consumers you can't have bandwidth-based pricing. 

<jeremy>We're mostly talking about the mobile market, and the drive to paid content for services rather than raw data as the unit for pricing/billing.  The discussion could have benefitted from this clarification, as many people think we're talking about broadband suppliers charging for access to third-party apps and content</jeremy>

Rick Rashid -- SVP Research at Microsoft

It's been 10 years since Mosaic.  It was really a tool for people to interact with computers at remote locations.  Driving change will be computers interacting with computers, and then servicing the consumer.  So with broadband in the home, the computers need to be intelligent and able to get data and services on the network all the time.

Today TV is more like browsing and you get data on the fly.  Increasingly, machines in the home wil pull content over the network and assemble programming for you on the fly, when you want it.  We've seen this ourselves in our own research environments.  For example, TerraServer went from being a browser app to a data service, and more and more its apps communicating with the the server for data.

Effecitvely, what the user sees are specific apps/services that are local on their machines, with communication and data happening in the background using XML web services.  So this is the real change and trend, this new distributed computing infrastructure.  Next is defining how we do trust, communication, discovery, and so on.

Question:  value of the data is the real deal, but the communications linkages are not the value?

Yes, that's right.  You're seeing very low costs for data communications --- dark fiber is getting dirt cheap.  What's the real value?  It's in what you do with the data and services around it.

Question: doesn't the web services strategy, with horizontal services, wont they also be commoditized?

Yes, that's right.  You'l get to a point...what is the specific value of the info/service you're providing?  If it's unique you can drive some revenue from it.  In many cases, if anyone can do it, then it will be a commodity.  Web services does help this commoditization.  The other side of the coin is that it enables new kinds of apps to be built.  People are building data federations --- linking past databases using web services, creating a facade of data services with higher-order value (e.g. aggregation and re-composition of services).

Question:  how does one maintain their brand identity in the web services space?

eBay (Jeff Huber) -- right now its additive, not detracting.  This is mostly on their supply-side versus consumer side.

Question:  are consumers comfortable with computers acting on their behalf?

Paul Florack (VeriSign) It is a key issue, and trust and authentication are crucial for consumer acceptance.  We'll leverage certificates for verification across a number of fronts involving web services.

Rick -- we've been concerned about this notion of there being no one place that can be trusted.  consumers will trust some things and not others, and thus defining web services security, is the need for multiple forms of certification and proof, and the idea of federation of trust, so you don't have to rely on a single trust mechanism.


2:35:39 PM    comment []

We have a distinguished group of panelists.

Russ Seigelman -- Kleiner-Perkins, ran MSN for Microsoft
Bob Meyers -- President of CNBC
Salvador Arias -- IBM consulting, communications practice
Rob Glaser (via video) -- RealNetworks CEO
Eric Schmidt (via video) -- CEO of Google

Thoughts from each panelist.

Russ -- big believers in broadband.  what happens next?  hard to know.  what's going to happen to usage?  broadband drives more use of the Internet.  believes that 1:1 video calls will be a big growth driver, will happen in mobile and at home.  business model of broadband is different --- lots of infrastructure investment will mean that there will be more tolls in front of the consumer.  it's the only way this will happen.

Eric -- the industry has been in a disaster for 2-3 years.  the issue has to do with the fact that computers are getting more powerful, and the value equation hasn't kept up with PC capability.  great thing about broadband is that it's the only technology that can meet the capacity of PC/CPUs.  mostly broadband has to do with persistence of connection (always on aspect) means a very different usage.  last year world switched from analog to digital devices in the consumer world.  new generation of digital devices, combined with broadband persistent connection, is the killer app.  vast majority of broadband adoption will be non-traditional, self-published audio and video.  result will be new works of art.

Bob -- work at CNBC, part of NBC, part of GE.  seek out environments where there are new challenges.  challenge of working on point-to-point, interactive versus broadcast world.  who's gonna pay for Internet delivered content?  who's gonna make the transition from traditional media into networked, IP broadband environment?  in the old day you have a dumb receiver (TV), receiving from a dumb but efficient network (VHF, Cable).  now the receiver and the network have more intelligence.  and the programming itself has a lot of intelligence.

Rob -- want to add to the comments already discussed.  Real's focus is on delivering a/v apps on the Internet.  started with narrowband, but focus was to design for the broadband world.  our sense is that in the mid-term, the medium will have the same transformative impact that cable had over broadcast, in terms of impact on business and programming models.  Real sees growth from dozens of channels that are profitable, to tens of thousands of profitable new channels.  Result will be increase in consumption. There will be a substitution effect.  Global medium, no channel or spectrum constraints, spans consumer and business equally.  Takes issue with Eric's idea that it won't be traditional programming/video that has distribution.  Real is now broadcasting 50 MLB games live over the network today, way more capacity than any cable provider.  ABC news is now doing 24/7 live news, all delivered over the Internet.  This is meaningful from a traditional media distribution perspective.

Salvador -- still focused on what's happening with the pipes and the economics of consumer adoption.  four key factors:  1) competitive dynamics b/w suppliers are driving prices/offers down, 2) need to have enough bandwidth to enable compelling apps (e.g. above 1.5MBs, esp video centric apps), 3) pricing elasticity having an impact, lower cost structures helping, higher demand equates to economies of scale 4) government intervention --- what will happen with government support to drive adoption; still uncertain, mostly industry driven now.  momentum is building for government supported investment protection (e.g. tax breaks, subsidies, etc.).

My question --- original programming, what will it be, will new companies emerge.

Rob.  Lots of tries during dot com era, and most tries were too early. Will play like cable tv, like early days when MTV and others started with low-cost programming.  Even ESPN and CNN were very parsimonious with cost structure.  Kinds of programming --- number one thing that you can do that you couldn't do in the past, is to add interactivity as an intrinsic component to the video.  Customization of programming, personalized information tools.  Other applications include community integration and real-time communications around the programming.  Social environment in addition to a programmed environment.

Bob.  Do you want to make it a business?  It's hard to make a business out of a much smaller niche service than we have today with large-scale media already.  As far as adding elements of interactivity, to realy create a new channel that will be video based, person-to-person; not sure how to do it, how to find it, but he can't see a way to a breakthrough that this will be a new business for them.

Russ.  There are now cable channels that are connected to live events on the Internet --- polling, video, community interaction.

Question: how much control will consumers have over timing of viewing of programming?  hard to imagine Tivo'ing CNBC, he says. need to watch it live.  says the key for broadcast is getting people to want to see it live.

Rob.  RealityTV interesting use case.  Doesn't have script format like other programming; it's the environment where media companies are combating traditional programming constraints; aggresive integration of marketing and advertising into programming.  Media companies will integrate marketing/ads into programming to combat Tivo style recording where you can skip advertisements.

Russ.  Time-shifting will work and people will pay for it.

Question:  what will happen with paid content?

Rob.  We jumped into paid content big time.  As advertising online collaposed, broadband grew, so the only business model became paid content, with a set of free content around it. Pragmatic approach today.  900,000 paid subscribers with a broad range of paid content.  Free consumption is still there and promotion and advertising supported, and its at an all time high in use.  These two models reinforce one another, because the free model creates an audience, which in turns drives more paid premium content, which in turns draws more people to the medium.  A la carte in music is just about to happen, its the play to save the music industry. 

Eric.  He disagrees with a few things. Says he's a huge fan of CNBC.  Majority of broadband entertainment will be ad supported media.  The advertising model is broken because of lack of ability to do sophisticated targetting.  Tivo model enables the ability to interlace targetted advertisements.  Online delivery in general supports this.  High-levels of personalization will mean a fragmenting of audiences, with a deeper mixture of blended advertising, marketing and programming (as Rob suggests).  Sees a huge change brought on by precision in advertising, and this is the revolutionary change on broadcast.

Question:  cost of technology and running a media system is still extremely high, much higher than even labor costs.  Couldn't break even on their projects. 

Rob.  Two different cost drivers.  One thing you saw in the bubble was that there was lots of capital, and so people were used to spending tons of new capital.  Real change since then and people aren't willing to do this so are looking for outsourced usage because its more cost efficient for the broadcaster/supplier.  Second thing is the rapid drop in bandwidth cost.  This variable cost is getting cheaper and cheaper, dropping 50% per year.  And it's accelerating.  So things like video on demand are now becomming feasible.

Russ.  It's true at the backbone, but not at the end-points.  This is still too costly and for not enough speed.

Rob.  The guys going bankrupt are the CLECs, but the ILEC/encumbant, they're doing fine, and actually are growing and profitable.  Those price points are a chokepoint on growth.  This is a glass is 2/3 full situation.  Broadband growth is happening.  Cable guys are nailing their numbers every quarter.  They're not seing a saturation at $50/month.  We'd love to see $20/month and the dramatic growth that would happen with that.  But this is a temporary inhibitor, and we're still seeing strong growth. Government should have a policy to drive prices down with subsidies, and yes, two player oligopolies don't price compete.

Question:  broadband is turning the communications and media industry inside out. And that it's not having a lot of relevance for corporations? 

IBM guy is talking about how all is good...not sure what he's really talking about....trying to incorporate IBM OnDemand marketing pitch...not connected....

Question:  we're really dealing with duopoly. isn't the real issue price?  won't there be real price competition.

Russ.  The real competition will be from wireless for the last-mile, perhaps powerline.  Real question is whether there is a price umbrella that will drive demand?  We're doing well now in terms of adoption, but it's still a long way before we get to 70% penetration.  Some scenarios could come in and really change the model, such as WiFi mesh networks.  Mostly theory now, but could happen quickly with real business model innovation.

Eric.  If you look at the adoption rate for tech as a proxy for the future, one of the fastest adoption curves is 802.11b/g.  Everyone is embedding this uniformly --- MANs, corporations, homes, etc.  Still some sharing, billing issues with mesh networks. It's pretty clear that the change to add the third wire will be distruptive, and will affect the pricing model in unanticipated ways. The way to get it down under $10 will be by users bearing some of the capital costs, and then participating in the mesh network.

Rob.  These new environments become real new platforms.  For example, I picked up a Cisco ATA router and have a VoIP phone device, and it really works.  This is one of a number of apps that switch over to the broadband pipe.  Big fields like education and health care could switch over to this medium.

Salvador.  Telco thinks, if I got to an IP based platform, can I really simplify my cost structure positively.  Lots of investment in their back haul and there's a concern about a displacment impact.

Question: the conversation has been around innovation around broadband, and talking about continuous video streams.  the Internet showed that a very servicable pipe happened when the browser happened and people had power at the end-point, and the innovation was around what you could do at the end-point, not the pipe itself.  real innovation is in the richer combination of media, communications, interactivity at the end-point.

Eric.  There are a couple of historic transformations.  The cost of published information --- any media type --- have dropped dramatically because of the digital revolution.  How do we get more useful media?  useful is in the eye of the beholder, and there will be thousands of capabilities.  Blogging is the other trend, which is a constant stream of thousands of perspective.  The Internet model gives the end-user more power and this is the real transformative impact on the broadband media world.  On another note, people are using broadband to transfer to physical media -- e.g. CD and DVD burning. 

Rob.  DVD and CD burning, etc.  The role that broadband plays in catalyzing media delivery in and around the home.  Once you have it in the home, it gives the user much more control to redistribute to fixed media, to devices, to get it to other fixed end-points (e.g. stereo and TV).

Question:  how many users will we need for the dramatic impact to society?

Critical mass will be 50% of households.

Question:  how do you se this playing out in a nice near-term number --- say 2008?  what will we be saying?

Russ.  It depends on what your expectations are for the outcome.  Maybe I'm just a dreamer.  I want to walk around everywhere and have a video phone and have it be high-quality and everywhere.  We'll be there after 2010.  It must mean that the majority of people have that.  It's not a 2008 kind of thing.

Bob.  It doens't matter.  We've reached a tipping point.  The comparison between TV to Cable change isn't appropriate.  We're talking about a communications medium, a revolution, becuase it's two-way.  The things people do on broadband they already do on narrowband.  Broadband just lets me do it better.  Nothing radical happens in the next five years, we'll just keep evolving and the network effect will take hold.

Eric.  In 2008, we'll all live in a parallel world.  At the same time, a subset of us will have new digital device with all of the worlds information available immediately, on the device.  I think there will be many thousands of chanells, derivative of what we see today, but highly personalized, and the vast majority will be small entities, with small audiences.  Much more information and end-user empowerment.

Salvador.  2008 the cable world will have significant improvements, such as one network that does IP, video, voice, etc.  It will clearly wipe-out dial-up and a broadband IP majority in the US.  It's not going to be video as we think of it, lots of community oriented stuff, real-time communications, gaming oriented content.  The world will be more on demand for us.  Broadband will get there.

Rob.  There will be multi-channel access to video and programming.  He shares Eric's vision of the portable revolution, with highly reliable, caching enabled devices, driven off WiFi hotspots.  Financial markets will see a great come back because of the inevitable progress of the Internet.


12:30:04 PM    comment []

I just wanted to plug the forthcoming Macromedia-focused developer conference in Europe, CF_Europe. I'm not sure why they didn't change the conference name, as the speakers and agenda are really very broad on all MX products, with a solid focus on building RIAs. Sister conference MXNorth in Canada was great, so I expect this to be as good.
10:27:06 AM    comment []

Rob Austin did a bunch of research and interviews with the creators of modern computing, like Bob Taylor, Alan Kay, Doug Englebart and Bob Metcalfe, among others.  Here's some of the things he learned.

Bob Taylor is one of the early innovators in networking technologies and the Internet.  Rob Austin, conference chair, talks through several examples from the late 1960's when Bob brought the idea of packet switched wide-area-networks (Internet), and IBM, ATT, and Xerox all passed.  What would have happened had they adopted the idea back then?

Bob Taylor:  "people tell me that the Internet happened fast; that's crasy, it took forever."

Some people can tell the future, gives examples of Vannevar Bush (hypertext), JCR Licklider (Man-Machine Symbiosis), Bob Taylor (The computer as a communication device), and Doug Engelbart (mouse, gui, real-time communications and collaboration).  These were ideas developed in the 40s, 50s, and 60s.

He reviews some interesting quotes from his interviews, and interestingly nearly all of them share some things:  they all are driven by the desire to help humanity; they all don't think we've nearly reached their vision, despite the progress that has been made since the 1960s.

Discusses the tension between invention/vision and short-term business objectives.  There is an internal tension in business firms that attempt to reduce long-term vision and R&D because they way in which it can cannabilize current products, prices and profits.  Cites Veblen (19th century economist) and John Kenneth Galbraith (contemporary nobel winning economist, advocate of "great society" economic policies under JFK).



9:51:13 AM    comment []

I'm attending the Broadband Explosion conference hosted by Harvard Business School. It's a great line-up of speakers and keynotes, and a small intimate crowd.  Rob Glaser, Sky Dayton, Russ Seigelman, Kevin Werbach, Geoffrey More, Clayton Christianson, Rich Rashid, Eric Schmidt, and lots of other good folks.

Today's opening keynote is Doug Van Houweling, President and CEO of Internet2, the government sponsored research initiative to create an NG-Internet.

He's giving a little history on the emergence of the commercial Internet, via NSFNet and a partnership with IBM, MCI and the state of michigan.  I remember this very well, and knew the folks running policy at NSFNet.  It was clear back in 1993 that this was going to be turned into a commercial backbone, and the rest is history.  But now the same model (government/commercial/higher-ed collaboration) is being used to create a next-generation Internet -- explosive broadband.

He's recalling the early strategic decision making that went on with IBM around NSFNet, and the controversy it had with IBM, who were marketing SNA networks as an alternate WAN topology.  At the same time they had built core router technology for this early Internet, before Cisco even had backbone routing technology.  IBM decided not to productize their backbone routers, and this opened the door to Cisco dominating the backbone routing infrastructure.

Internet 2 Charter: enable new applications, re-create R&E network, transfer technology to production Internet.   It's recreating the NSFNet model, etc.

202 universities have joined the Internet 2 project, 50 corporate members, 30 government and research labs, and over 30 international partners.

Compares time to download The Matrix DVD using various connectivity tech.  56k model is 171 hours, DSL is 25 hours, T1 is 6.4 hours, and on tests on Internet2 it's 30 seconds.  Sounds good to me. :)

Blasts today's broadband.  "Real Broadband", as Marc Canter would say.  Multi-megabit bidirectional, and services rich -- e.g. multicast, IPv6, deep facilities for measurement, monitoring, metering.

What current experimental applications are being deployed on this today?

  • Media distribution --- away from sattelite to IP digital distribution; HDTV over broadband.  Super quality video.  Experiments with 4x quality of HDTV.
  • Telepresence.  Extremely high-quality remote presence and monitoring.  For example, underwater observatory with mobile monitoring, cameras, etc.
  • Robotics and bio-feedback.  Devices that record behavior -- e.g. a surgeon working with instruments, and end-users who can re-experience that movement with simulated devices, wearing VR goggles.
  • Real-time collaboration.  High-quality multi-location video conferencing with screens the size of whiteboards.

The infrastructure needs must change dramatically. Today's Internet is "best efforts" Internet -- e.g. make a best attempt to get data there quickly. Internet 2 is guaranteed level of quality, which translates to a much larger set of pipes, routers, etc., but more importantly is a services infrastructure in the network geared towards quality of service such as IPv6 and multicast.

Also believes Internet 2 needs to embrace federated policy models for security and trust.  They aim to build proof-of-concepts for security models that work across enterprises.  They're working with the Liberty Alliance and Microsoft, and trying to get something working today in the Internet 2 architecture.


Biggest challenge is the last mile, given the role of the end user in distruptive change and value on the Internet.  Internet 2 can enable Gigabit to the home, but this requires a partership between government, community and commercial service providers.

9:11:41 AM    comment []

© Copyright 2004 Jeremy Allaire.
April 2003
Sun Mon Tue Wed Thu Fri Sat
    1 2 3 4 5
6 7 8 9 10 11 12
13 14 15 16 17 18 19
20 21 22 23 24 25 26
27 28 29 30      
Mar   May

Click here to visit the Radio UserLand website.

Subscribe to "Jeremy Allaire's Radio " in Radio UserLand.

Click to see the XML version of this web page.