Communications

Google and China — a very different view

Usman Latif (Techuser) has a fascinating take on what’s going on with Google and China.  I can’t guess if he’s right or not, but we’ll know in due course.  If what he suggests is true, Google will exit China by selling it’s Chinese business to the Alibaba Group.

In 2005, Techuser uncovered the strange nature of Google’s 2004 patent settlement with Yahoo.  The terms of that deal have never been revealed, but what was revealed in various SEC filings leading up to Google's IPO appears to have been, at least, obscure and misleading.  Usman’s point now is that 2004 deal included some substantial, probably very substantial, undisclosed obligations, perhaps due at the end of an unknown grace period.  So Google may need a way to make a big payoff to Yahoo without letting anyone understand the details of their patent settlement.  Turning over their rapidly growing Chinese business to Alibaba (40% owned by Yahoo) might be their answer.

Yes, it reads like a “conspiracy theory,” but Usman is a smart guy and, in any event, we soon know if he’s right or off track.

Walled gardens have collapsed

Yesterday, Mary Meeker of Morgan Stanley gave her annual state of the economy and the Internet presentation at O'Reilly's Web 2.0 conference.  I wasn't there but her presentation is on line and it has some interesting insights.  Many others are commenting.  The one thing I'd like to draw to your attention is slide 46.

In one year (2007 -> 2008), UK users abandoned their wireless carriers' portals and went directly to numerous open mobile internet services.  I'd love to see comparable data for the US.

Google's Peering and Caching Strategy

This is a followup on yesterday's post about how little Google/YouTube pays for bandwidth. Google wants to promote peering with ISPs, so they give presentations at ISP meetings.  After reading yesterday's post, Alex Benik of Battery Ventures sent me a link to this presentation given by Google at a 2008 meeting of the Latin America and Carribean Internet Addresses Registry (LACNIC).

As expected, Google peers with as many relevant ISPs as possible.  For the ISP, peering with Google eliminates their upstream costs for traffic to Google.  Since Google represents a substantial volume of traffic for most ISPs, this is a big saving.   As of May 2008, Google was present in 33 public Internet exchanges around the globe, so major ISPs already have connections in places where they can peer with Google.  The minimum qualifications are 5 Mbps of Google traffic and the ability to interconnect using Gigabit Ethernet at one of these 33 major Internet exchange points.

Google Global Cache

What's interesting is Google's caching strategy.  Just as Akamai puts servers in ISP's local facilities, Google is providing a distributed cache for their content.  This available to larger ISPs and allows them to serve Google content directly at the edge of their networks, thus reducing traffic on the ISP's backbone network.  Here's a representative rack that Google provides to the ISP.

Can You See Me Now? Video-enabled call centers...

Some weeks ago, CRM magazine asked for an article on video-enabled call centers.  This idea is a bit futuristic for the US market, but such call centers are actually showing up in Asia and the EU, at least experimentally.

Think of calling a help line and being asked to point your video telephony handset at the control panel of the appliance that's causing problems.  That's becoming possible in some Asian and EU markets where most 3G handsets support video telephony and 3G penetration is well over 50% (and much higher in Japan).

In any event the article,

Can You See Me Now? 
Video-enabled call centers will change the company-customer relationship

is up on the DestinationCRM website.

African telecom: corruption mostly limited to publicly run operators

The latest issue (#421) of Balancing Act News Update has an interesting story based on the prosecution (in the US) of former sales agents for ITXC found guilty of bribery in arranging VoIP connectivity contracts.

Six of the seven telcos involved were state owned businesses and in the one with private ownership (Sonatel, the PTT in Senegal that's partially owned by France Telecom), it was France Telecom that got suspicious.  Balancing Act News goes on to point out the problems with bribery:

It will be clear from the summary of the available public documents that this approach to business has a number of pitfalls. The bribed employees in two instances subsequently had to sort out “cost disputes” which were the pretext for asking for more bribes. In one instance, the bribed employee failed to share his gains with the managing director. In the only company with a private shareholder (which also sells wholesale minutes), the bribed employee became isolated under commercial questioning: this provides a strong if not always decisive argument for privatisation.

Based on their conclusion, it appears there is no viable way to do business with the state controlled Telcos in Africa:

From the telco side, what has been known privately for years is now revealed clearly and publicly by the Rwandatel and Sotelma examples. Corruption is systemic and goes right to the top and bribes are expected to be shared with the boss of the organisation and woe betide anyone who tries to keep the money to themselves.

Luckily, the mobile telecom industry in Africa is mostly private and increasingly competitive, with the result that mobile phone adoption is soaring in Africa.  It would be nice to think the PTTs will be cleaned up, but in any event, the benefits of telecom are reaching African citizens via mobile phone services.

One patent justifies Microsoft's interest in Yahoo

I've watched coverage of Microsoft's bid for Yahoo! and the related maneuvering between Google and Yahoo!.  The explanations are not very convincing.  Microsoft doesn't need Yahoo's search technology or their morale-impacted work force.  Yahoo's search market share continues to decline and there's little of strategic relevance in the rest of their business.  What's the attraction?

Usman Latif has the first interesting explanation I've seen, which he explores in depth in a ten page article at Techuser.  The issue he uncovers is US patent 6,269,361 on the bid-for-position paid-for search mechanism that funds Google and indeed all of the Web 2.0 phenomena.  Patent '361 was issued to GoTo.com, later called Overture, in 2001.  Yahoo! acquired Overture in 2003 and used the '361 patent in litigation with Google shortly thereafter.

Usman is uniquely positioned to recognize the importance of this patent as he has followed paid-for search patents for years and has previously written (in 2005) about the strange settlement between Google and Yahoo, reached on the eve of Google's IPO.  Google negotiated a perpetual, royalty-free license to this patent, but it is not irrevocable!  Usman's surmise is that Google's license would be revoked if they sue Yahoo over any other intellectual property.  That's a commonly requested clause in such agreements and, on the eve of their IPO, one that Google likely had to submit to.

If you are at all interested in patents or in the current Microsoft-Yahoo-Google machinations, take a look at Usman's paper.  It makes more sense that all the fluff that's been written in the financial press.

Another analyst under estimates mobile subscriber growth

From a recent press release

Worldwide mobile subscriptions will rise from 3.9 billion in 2008 to 5.6 billion in 2013, according to a new Strategy Analytics report.

Sorry guys, you're forecast has fallen into the same trap as almost all the five year forecasts I've seen since 1995.  About the only credible analyst's statement in this area is one by Mark Newman, head of research at Informa, who said (in 2007):

The mobile industry has constantly outperformed even the most optimistic forecasts for subscriber growth.

The problem is analysts' forecasts are consistently pessimistic.  Consider Strategy Analytics' numbers.  They project 1.7 billion new subscribers in 5 years (60 months).  By every measurement I've seen, mobile subscription growth has been over 50M per month at least since 2004-2005.  For example, 2 billion subscriptions in September 2005 and 3.3 billion subscriptions in November 2007 comes out at 52 million new subscriptions per month.  If you believe Strategy Analytics' forecast of 3.9 billion by the end of 2008, that's 54 million more per month for the balance of 2008.  If we merely sustain current rates, we will hit Strategy Analytics' December 2013 forecast in just 31 months, i.e. in July 2011.  So they expect current growth rates to decline rather substantially in coming years. 

Why would they think this?  I haven't spoken with anyone at Strategy Analytics, but in discussions with other analysts over the past ten+ years, a frequent answer is "current growth rates will have to slow as we're running out of people who can afford mobile phone service."  Wrong!  This is the view of someone who doesn't understand Moore's law.  Whether it's transistor density or wireless performance, innovation drives exponential improvements in price-performance.  Per-capita GDP is improving slowly, but the cost of mobile phone infrastructure and mobile handsets is dropping substantially every year.  Today, those at the bottom of the pyramid are in a position to at least use a phone and increasingly to acquire their own mobile phone.

In addition to innovation ― at work since the introduction of mobile services ― political obstacles are falling in many countries.  Since 2000 it's become clear, mobile phone adoption brings significant economic advantages, and the best way to get mobile phone adoption is to attract multiple competitive mobile phone operators to your country.  Increasingly, even the most backward dictators are realizing there's more money to be made taxing multiple competitive mobile phone services than in keeping all the profits from one government phone company.  So country after country are encouraging competitive mobile phone operators and allowing substantial foreign investment.  This political trend serves to increase mobile phone adoption rates.

Of course, adoption rates will saturate at some point.  When might that be?  The best evidence we have is for 2G and 2.5G services in developed countries.  It appears things slow down a bit around 120 subscriptions per 100 people.  Of course this may change with 3G.  It's too early to tell.

Technology adoption generally follow an S-curve:

If saturation is at 120 or above, and we're currently around 50, we have at least another five years of rapid growth.  Perhaps a five year forecast made in 2013 will be justified in projecting a decrease in adoption rates.  But by then we'll also understand the impact of 3G.  Who knows? 3G may drive things even faster. 

Or perhaps by 2013, I will have a single mobile data subscription for a gateway device on my body, that in turn provides connectivity for all my other devices.

In any event, I expect 5.6 billion mobile phone subscriptions by the summer of 2011 and more than 6 billion by mid 2012.

Mobile video-on-demand Yes! Mobile broadcast TV not so hot

I have an article, Going Mobile (TV), that's recently been published by MobileIN, a wireless and mobile information site.  In it I basically argue that major investments in mobile TV broadcast capability are less likely to pay off than investments mobile video-on-demand.

The biggest trend in commercial television viewing is personal video recorders like TiVo.  People want to watch TV content when they want, not when broadcasters schedule it.  The only exception is major sports events (the Superbowl or World Cup matches).  Even the evening news is frequently rescheduled for later in the evening.

The second relevant trend is growth in YouTube and similar web-based video content.  Broadcast TV went from 2-3 channels in the 1950s to hundreds of channels on a typical cable system today.  But consumers are also interested in the long tail of millions of videos that can only be served over the Internet today and, potentially, over the mobile Internet in the future.

Finally, survey's of early adopters of mobile video show music videos, movie trailers, weather, sports action clips, comedy videos, cartoons and amateur video shorts – typically a few minutes long at most – are the most popular content. In addition, it appears 85 percent of mobile video viewers watched viral videos (content sent or pointed out by others) rather than content they found themselves.

All and all, mobile consumers are looking for video -on-demand, not pre-scheduled broadcast TV.

So what's the logic for massive investments in spectrum, followed by even more money in new wireless infrastructure, followed by the need to sell everyone new handsets that can receive the new broadcast mobile TV channels?

The end of Yahoo, and the end of Microsoft

If you are not familiar with Umair Haque &/or don't regularly read his Bubble Generation blog, read Umair's views on this potential merger. 

He begins with "Yahoo + Microsoft isn't just a mistake - it's a double suicide."

It gets better. 

And his argument is independent of the culture clash between the these two companies (a subject eloquently raised by his first commenter).

How tiered Internet pricing could actually facilitate P2P

Time Warner Cable's planned experiment with tiered charging for Internet access has generated a flurry of coverage in the blogsphere, but no new insights (at least that I've seen).

The primary problem ISP's complain about is that 5% of their customers use 90% of the available bandwidth and when they examine this traffic, it's mostly peer-to-peer file sharing.  A reasonable question is how to allow as much of this traffic as possible without increasing an ISP's variable costs or slowing down their other users.

This may not be as difficult as it appears.   Indeed if Internet access was as competitive as mobile telephony, we might already have seen what I'm about to propose — a combination of bundled pricing equivalent to mobile's "free nights and weekends" and "free on-net calls" with a way to facilitate P2P traffic that leverages exactly these "free" periods.

An ISP's costs

ISPs have some costs which are relatively fixed and others that are tied to usage.  A network is a relatively fixed cost and when it's not full, the incremental cost of adding traffic is zero!  This is the reason mobile operators give away free nights and weekend.  They've built their mobile network for the peak daytime traffic, so it costs them nothing to run promotions that add incremental traffic at off hours.  Peak hours and off hours may be different for an ISP, but the concept is the same. When a data pipe is lightly loaded the ISP's cost of adding incremental traffic is zero.

On the other hand, some ISP costs are usage based, for example "IP Transit" or more properly, Internet Transit.  This is the ISP's upstream cost to send and receive traffic to/from the rest of the Internet.  However, even here, usage-based costs occur at heavy usage.  Light usage periods don't save money.  To understand what's happening, it's worth a digression on Internet Transit.

Internet Transit

Internet access is monopoly or duopoly or a heavily regulated industry.  The middle mile connections from the local network to the Internet backbone may or may not be competitive depending on where you are.  But the Internet backbone itself is extremely competitive.  If you can get to a major Internet Exchange Point in the US or Europe, there are many providers offering extremely competitive rates for Internet Transit.  Typically these services are priced on a megabit per second per month basis (Mbit/s/Month) with lower rates for higher volume commitments.  The other key idea is that charges are based on the 95th percentile of all the five minute data rate samples taken during the month.  So an ISP can have a few bursts above their typical rate, as long as they represent less than 5% of the sampled intervals.

But this also means there is no extra cost to run at or near the typical rate at all times.

Local traffic

Even more important, if file sharing is done with other computers on the same ISP's network, then there is no need to pay for Internet Transit at all.  The question is how to figure out which potential peers are "on-net" and which are "off-net."

Sending signals to P2P software

Most P2P file sharing software has relatively little knowledge of locality.  Some P2P software practices "prefix awareness," for example, Joost gives preference to peers in the same /24 IP address block when they are available.  But if a major operator provided an automatic way for P2P client software to determine whether a prospective peer's IP address was currently reachable "for free", it seems likely the file sharing community would leap on it, and if there's money to be saved, active file sharers would download the new clients immediately.

A standard way to present such information might be via an extension to the XML-based response codes in one of the whois information exchange proposals, e.g. from ICANN or from APNIC.  Also, while what I'm proposing might start as a pricing plan rather like a mobile operator's "free nights and weekends" and "free on-net calling," it's not hard to see extensions where an ISP could offer dynamic access to underused capacity to those programs that were prepared regularly interrogate an ISP's server and use just the advertised off-hours capacity.

In closing

People liked fixed price deals.  Unlimited is great, but there's plenty of experience with bundles of minutes and the idea of data bundles has already showed up in 3G mobile data plans.  The combination of several tiered data bundle prices with the availability of "free" connectivity for "on-net" peers and during off peak intervals is likely to appeal to file sharers and produce better results for both the sponsoring ISPs and file sharers alike.

Global mobile network operators subscriber data

There's an interesting free source of market data that I've been using for the past two years.  Until two hours ago, I hadn't given this much thought, but when I mentioned it to some colleagues, their reaction was:  Wow!  This is really cool. 

So I thought I'd share it more widely. It's Wikipedia's List of Mobile Network Operators.

 

There's quite a bit of information, by country and by operator, for example:

I've been tracking mobile subscriber growth for more than ten years, in part by purchasing ITU data every other year or so; in part by getting copies of any data than any NMS business unit purchases from traditional market analysts and in part by capturing anything that's published in press releases or other teasers from analyst firms.

Wikipedia doesn't have all data for all operators, but they have almost everything and it's very credible.  They also cover countries that I have never seen in traditional analyst reports, for example:

Somolia is famous for having no functioning central government, no spectrum regulation and no telecom regulations, but none-the-less having the lowest international calling rates in Africa.

Ringback Tones Adoption Rates — Culture, Biz Model or Marketing Savvy?

Ringback tone services have been wildly successful in many parts of Asia and they are emerging as a major money maker, not to mention a major music distribution channel, in many other countries, but not in the US or the EU.

Nearly 55% of Korean subscribers have ringback tone service.  China Mobile has close to 50% adoption.  Yet, as of September, adoption rates in most of Europe were below 10% (the UK was at 2%) and US adoption was only 5%.  Since our LiveWire Mobile division provides white label managed ringback tone services and sells ringback tone service platforms, I'm obviously interested in understanding more.  And, with 30+ operators on six continents using LiveWire Mobile platforms or services, I have some access to comparative data.

A late start is only part of the story.  Korea went from 0 to 30% penetration in just nine months (April to December, 2002).  China took a bit longer as the service only became available in stages across China Mobile's 31 separate, relatively autonomous, provincial units.  That roll out took from mid-2003 to early 2004.  But by the end of 2004, China Mobile had 30 million ringback tone subscribers (then 69M by 12/05 and 144M by 8/06).  And yet after ~2 years of service, the US has only 12M subscribers or 5% penetration.

One clear difference in Korea and China, versus the US and EU, is the number of independent marketing organizations directly or indirectly promoting the new service.  I remember visiting Seoul Korea in January 2003 and having a billboard pointed out to me where a Korean rock star was advertising his latest hit and how you could make it your color ring tone. I was told there were more than 20 independent organizations providing content for SKT's Color Ring service at that time.  More significantly, revenue splits were such that these independent organizations found it worth while directly promoting their content to mobile subscribers.  That means there were 20+ marketing teams with independent activities that all served to increase subscriber awareness.

Compare this with the typical operator in the US or EU where the only source of information or content is the operator's portal and the only service promotion activities are those of the operator.  That means there is one product manager focused on ringback tones, they have a limited budget and they must compete with their associates for position on the operator's WAP deck.  Is it any wonder awareness is low?

Of course other factors also contribute to service success.  Our LiveWire Mobile team has accumulated a set of best practices based on discussions and consulting work with operators around the world.  They have a formal service offer of course, but I'll endeavor to summarize some of their learning in a future blog post.

LiveWire Mobile — new brand for well established business

This morning, NMS Communications launched LiveWire Mobile as a new brand for our mobile applications business.  I'm leery of re-branding exercises but this was long overdue as our mobile applications business is substantially different and independent of our traditional developer focused business.  Now LiveWire Mobile is operating as a distinct division of NMS with this new logo:

LiveWire Mobile focuses on mobile personalization services, including our well established ringback tones business.  That makes LiveWire Mobile a market leader from inception, as our ringback tone service is deployed with over 30 operators around the world.  The most recently new operator announcement also came today — it's Virgin Mobile USA.

Mobile Personalization

Mobile personalization services hit some years ago with ringtones and wall papers.  Many think of ringtones as a content business, and yes, there is a content sale in many cases.  But whether it's ringtones or ringback tones, the key motivation is the human desire to personalize our possessions and our environment. 

Today, ringtones are widespread and revenue growth is slowing.  However, ringback tones are still in the early growth phase, at least in Europe and North America.  Ringback tone penetration is over 55% in Korea, but less than 10% in the US.

Besides the established ringback tones base, LiveWire Mobile has plans for additional network-based message and subscriber-focused personalization services — stay tuned.

UPDATED:  Here's the link to a press release with more info (in PR prose...).

Managed Storage Futures

Recently I wrote about differences in the exponential growth rates of computing and networking and promised to say more about how these differences cause substantial shifts in the technology landscape.  Managed storage is one example.  The relevant doubling rates (from that earlier post) are:

Doubling Rates

Technology Measure	Months
Computing performance	18
Storage capacity	12
Networking performance	15
Access connectivity	20-26

The increase in storage capacity per dollar has been phenomenal and is one of the reasons that Google can offer Gbytes of free storage for email and that Amazon can offer their Simple Storage Service (Amazon S3) at extremely low rates.

But it's also caused headaches for IT directors, as installed equipment becomes obsolete long before it's fully depreciated, and employees and department heads grip about inflated internal billing rates for storage.  Pity the IT staffer who sends a broadcast message justifying corporate email storage limits because "it costs the company X cents per megabyte per month."  I've seen such messages, and the employee ridicule they engender.

This sounds like a perfect opportunity for a managed service — provide an interface that looks a storage area network or network attached storage, using multiple (for reliability and arbitrage) Internet-based storage services to provide the actual storage.  But now differential growth rates become a factor.

Storage costs decline a bit more rapidly than the cost of Internet transit.  So, already it's the case that network-based storage is extremely low cost for backup but less affordable for transactions.

But the real problem is the cost of access connectivity.  If you're selling managed services to IT departments, you need to provide services at their premises.  Local connectivity is not fast, cheap or reliable, and the pace at which it improves is glacial in comparison with storage or Internet transit.

Has this prevented the emergence of managed storage solutions?  Of course not.  But most existing solutions focus on remotely managing equipment that's physically on the enterprise premises.

Is there opportunity for network-based managed services.  Also, yes.  But you will need considerable focus on local connectivity, both for the numbers you use in your business plan and for the specifics of how you implement the service.  Some thoughts:  interface your managed service via a remotely managed on-premise box that includes caching?  use a dedicated access link to guarantee QoS?  ???

In any event, three years from now, you can count of disk storage being ~8X more affordable, Internet transit being perhaps 4x more affordable, but local connectivity only 2x or 2.5x.  Don't give up your great  business idea, but plan accordingly.

Making a significant contribution -- Hamming's career advice

While following an entirely different thread, I stumbled on a blog post by Giacomo 'Peldi' Guilizzoni which lead me to the transcript of this really significant talk that Richard Hamming gave in March 1986.  I printed it out a few weeks ago and just read it today. 

Wow!

His title is "You and Your Research" and he discusses what it takes to do really great work and what are the differences, among otherwise smart people, that cause some to do great work and others to be forgotten.  Hamming was a mathematician (familiar to EEs for Hamming codes and to DSP engineers for Hamming windows) and this talk is about what he saw among scientific colleagues at Los Alamos, Bell Labs and the Navel Post Graduate School.  However, what he says is broadly applicable to any field of endeavor.

Read "You and Your Research."

Watching the leaks about Google's Open Handset Alliance

11:15am EST and the Open Handset Alliance website is (finally) live.

I'm subject to an NDA, so I've just watched the leaks without comment.  The Wall Street Journal carried the first news almost a week in advance, but without any mention of the name Open Handset Alliance.

I began intermittently checking the results of a Google search on the complete phrase "Open Handset Alliance" in anticipation of leaks in advance of the actual announcement.  Up through Friday, there were zero hits.  Likewise on Saturday morning, at least for a web search, but a Google news search returned the first hit.  It was this from CNET News at 6:25pm Pacific time on Friday evening.  From the text, they may have gotten their information from someone in Japan.

By Sunday midday, the conventional Google web search was returning this one article from Friday midnight (Pacific time) which credited the CNET News story from earlier that evening.  However, a Google news search gave the CNET article directly (as the 3rd item) with two more recent articles above it.  This situation remained stable through Sunday evening (eastern US time).  But by Sunday evening, there were 24 blog posts tagged "Open Handset Alliance" on Technorati.  They all traced back to the CNET News or an alternate copy of the same story.  Finally, sometime between 7pm and 9pm Sunday evening, Google must have updated their database as suddenly they were reporting 14,900 hits for the complete phrase "Open Handset Alliance."  By Monday morning this was 31,200.

Sunday also saw a New York Times background article – an in depth article about Andy Rubin (the lead for the OHA software project) that had obviously been set up with cooperation from Google.   As a cooperative venture, this article did not mention the term "Open Handset Alliance."

So across all the noise, it appears there were only two significant leaks:

1. On Tuesday the 30th, the Wall Street Journal said the announcement would come within two weeks and gave quite a few details.  Then on Thursday the 1st, they pinned it down to Monday the 5th.
2. On Friday the 2nd, CNET News broke the name "Open Handset Alliance."

Rather impressive secrecy for an effort involving more than 30 companies.

Solutions beat equipment

Bob Schechter, CEO Of NMS Communications, kicked off the Connect 2007 conference in China with a global overview of mobile markets.  One interesting point he mentioned was the terrible performance of the equipment provider segment despite rapid growth in mobile subscribers.

Here are the market values of some major communications equipment companies as prepared by Goldman Sachs a few weeks ago, i.e. before Ericsson went down 25%:

The rank here is Cisco, Nokia and then RIM! (equipment and service), Ericsson, Motorola, Alcatel-Lucent, etc.

If you look at the stock price performance of major equipment providers, it’s awful.  Here are the stock prices of Cisco and Nokia — neither has recovered the value they had the bubble that peaked in 2000.  Meanwhile solution companies like RIMM and Apple are far above their 2000 price levels.

Not a pretty picture for the equipment business.

 

Mashups: Web Meets Telco

Joel Hughes, General Manager for Mobile Applications at NMS, is moderating the second to last session at Connect 2007 entitled "Mashups:  Web Meets Telco" with panelists:

Matt Gross, Director of Product Management, WHERE by uLocate Communications -- a GPS-based platform for location aware services.  Other value they bring is that they have negotiated deals with several US operators (Alltel and Sprint) and they make these services available to 3rd parties.

Alan Quayle, Consultant, Business and Service Development -- talking his consulting practice and refers to British Telecom's Web21c as a sample application he's been involved with.

Kevin Nethercott, President and COO, LignUp -- talking about their web platform to program PBX functions and do enterprise mashups, specifically mentioning Mashup Camps and applications prototypes done in 15 minutes.

Sam Aparicio, CTO, Angel.com -- an IVR and call center solution that exposes a web API for control.  His example application is a mashup with Salesforce.com.

This real time blogging is new for me.  Perhaps the resulting notes will only be useful to me...  In any event, this is going to be more stream of consciousness...

First impression: there are several sources of telephony controlled by web interfaces, but they all cost money or are just sandboxes.  No one has mentioned APIs that let developers also participate in billing, or ways that web telephony can be available at or near free.  Surely one or the other is needed to see this industry take off.

For now, all the discussion is about enterprise developers.  Mashups are a new way for enterprises to develop their IVR and call center solutions.  So far, no examples of developers targeting consumers.  Alan has mentioned that BT's API has also been used for Salesforce.com integration.

Sam just alluded to a Verizon API.  I'm not sure what he's referring to.  I'll try and ask afterward...

Now Alan is discussing cost.  When location data cost 25 pence per inquiry, no one used it.  No one has a good suggestion of how these kind of services should be priced.  What about an application that needs two database dips, an international phone call and two international SMSs?  General agreement that this is an issue, but no specific suggestions or pointers to solutions.

The popular APIs to mashup with (besides Salesforce.com) are Google Maps, Flickr, Facebook and Amazon (for storage), at least today.

Now Joel has explicitly asked how things are priced.  Amazon is priced by the GByte per month and Salesforce.com is by user per month.  Enterprise telecom groups either buy the platform or pay per transaction until they buy the platform.  Their issues are reliability and minimizing costs.

Matt Gross describes the value uLocate brings to Sprint is to handle the long tail of developers that Sprint can't deal with directly.

Alan says BT's Web21c has helped expose issues and challenges.  They are still experimenting.  As far as location goes, the pricing is still too steep.  Alan expects that Google and others will note the GPS coordinates of every cell site, retrieve that from the handset and skip interfacing with the operator.  Operators will have to price their location info appropriately or they will be bypassed.  Consumers do seem willing to pay for safety-related location-based services.

Relation of mashups with IMS.  North American operators are adopting IMS to control VoIP telephony, but IMS is just the base for connection control.  The broader service delivery platform is less standard and yet that's what's required as the platform that supports web services.  SIP-based applications going a lot slower than web APIs.  In the end it's applications that matter.  No one on this panel cares about IMS except as an underlying layer that supports their web APIs.

A question from the audience: how do you see the handset interfaces w.r.t. mashups.  Kevin is focused on voice telephony.  He doesn't see the mobile handset as the application UI for now.  In the long term, he expects browser-based user interfaces, but for now, he's doing voice.  Alan seems to agree.  Handsets are so different that all you can rely on is basic voice and SMS.  Sam complains that even the so-called open API phones, you can't actually to deep features on the handset.  This crew is focused on voice.

New mobile applications, but how good are the business models?

There are four application vendors on the Applications Innovation panel at Connect 2007 in Boston.  These folks got 3 slides each and 5+ minutes to explain what they are about.

Colm Healy, CEO, Xiam -- mobile content discovery based on the automated discovery of consumer preferences.
Sunil Vemuri, Co-founder and CPO, QTech, Inc. -- Their product reQall helps people remember things (to do lists, names of people you meet, shopping lists, etc.) using a combination of mobile phones, web, WAP, SMS.
Nicolas Arauz, Co-founder and Managing Director, Xipto -- Endorsement-based mobile advertising where consumers choose to run advertising during the ringback tone interval, but only for brands and ads they are willing to endorse.  In exchange they get credits for the mobile service.
Karen Cambray, CFO, Groove Mobile -- Full track download pioneer.  Two million songs in your pocket.

Some immediate comments from the analyst and investor communities:

Seamus McAteer of M:Metrics -- the US is not the mobile backwater people usually say, at least for music services.  Subscription services are the only really viable models today.  Per transaction sales pale in comparison to monthly subscription revenues.  It's not rational (on the consumer's part), but it's what consumers want. 

Stan Reiss, General Partner, Matrix Partners -- it's still very early for mobile Internet.  So far there is not a good model for how to monetize the mobile Internet.  The operator's view that they are going to charge per transaction is very inhibiting (versus "free" on the Internet).  Carriers are making money and content folks are making money, but startups in the middle are being squeezed out.  Investors are looking for business models that can stand up to the carriers.

Strong counter from Nicholas at Xipto that you have to think about attention.  There's real (advertising) money in having a user's attention.  Of course Xipto is still in fairly low profile, but his argument included compelling figures for what advertisers might pay per ringback play (to one caller) compared with the content sale (music purchased a few time per year).

Seamus seems to agree that advertisers will pay substantially to reach people in any venue.  So there is a substantial opportunity in offering brands access to mobile subscribers.

Stan points out the cost of advertising on the web is extremely low.  The challenge with mobile is the carriers view of the cost of transferring content to the phone is still unreasonable.  Example, the US carriers are unwilling to do anything that costs less than a dollar.

Karen points out that Europe is ahead of the US in allowing off-portal or off-deck, i.e. wholesale, content sales.  The key is getting reasonable charges for data usage.  Some discussion of data charging models that are emerging in Europe, including capping the data fee per day.

Dean Bubley again had a good question from the audience:  most users in the world are prepaid, not subscription.  Both Seamus and especially Stan commented that it's easier to make money in markets w/o a lot of prepaid, if only because the consumers are wealthier.

Mobile Interent - a middle ground? for how long?

Some highlights from the first panel at Connect 2007 in Boston...

Andrew Budd of mBlox is very articulate on the need for wholesale models, i.e. not the "real" Internet, but a mobile world where operators allow (facilitate) customer access to 3rd party content. This is the compromise position between the typical walled garden operator and the completely open Internet (which Dean Bubley pressed in a question from the floor).  Andrew's argument against the wide open mobile Internet is roughly that we don't want the 900 number meltdown that we saw in the US in the 1990s (when premium rate services became associated with scams and porn).

Likewise, Jud Bowman of Motricity, clearly expects the "real" Internet to win eventually, but in a long transition in which operator run programs dominate.  Jud made an interesting point I need to double check:  PC's are only upgraded every 4-6 years (can it be that bad?) whereas mobiles are upgraded every 2 years (in the US).  The implication is we'll see very rapid change in the mobile market in the next 2-4 years as handset vendors learn from the UI of the iPhone and 3G data capacity continues to evolve.

Michael Scully, Director of Music, Mobile Content and Data, at Virgin Mobile USA was the one operator representative on the panel and, as might be expected, (and completely legitimately), pressed the point that Virgin Mobile has a relationship with their customers that centers around a very personal device - their mobile phone.

Another set of numbers (from Andrew Budd) that I need to check out:  The mobile industry is worth $700B whereas the Internet is only worth $150B.  That was partly contested by Seamus McAteer of M:Metrics but I didn't hear an alternate set of numbers.

Panel discussion, no slides, is a breath of fresh air

Listening to the first panel at Connect 2007 and my first reaction is Wow - No slides is the way to go!  Perhaps it's the quality of the speakers or Bob Schechter's moderation, but the nature of the discussion is much, much better than at a typical industry show. 

More on fat tails in stock prices

While it's a bit remote from communications, I've touched on mathematical models for financial markets several times (here, here & here), primarily as an outgrowth of my interest in long tails.  So I was very happy when Paul Kedrosky pointed to this an excellent survey by Dan diBartolomeo of Northfield Information Services.  The last eight slides are full of references – more than I'm likely to track down in a year...

There's a PDF version here.

Optimism in Africa

More breakfast reading..., of commentary generated during and after the recent TED Global 2007 conference in Arusha, Tanzania.  Technorati has pointers of course, but the most prolific blogger is Ethan Zuckerman with almost verbatim transcripts of each session.  And, the talk I found most interesting was that by former Nigerian finance minister Ngozi Okonjo-Iweala. 

Some excerpts from Ethan Zuckerman:

... we’re seeing changes in Africa that we never thought would happen. We’ve seen annual growth of 5%, in some cases 6-7%, up from 2%. External debt has been massively reduced. Countries are building up foreign exchange reserves, shoring up their currencies. Private investment flows are increasing, remittances to Nigeria are skyrocketing, and there’s a net inflow of capital.

But Africa needs jobs. 62% of Africa’s population is under 24. We have to figure out how to make these people productive. Nigeria is now building an opinion research organization, a way of listening to citizen voices, which she notes is a rare thing on the continent. The top issue in every survey? Jobs.

Then after a heart felt story of saving her sister's life with help from a clinic sponsored by foreign aid,

Okonjo-Iweala tells us she doesn’t believe aid, even aid to save lives, in the sole answer. We have to use it well. Why has southern Spain developed? On the back of aid which was provided to build road and infrastructure. Ireland is one of the fastest growing economies in the world - they used aid to build infrastructure to build an information society. “They didn’t say no to aid - but if they can build infrastructure in Spain, why do they refuse to build the same infrastructure in our countries?”

Finally,

The Chinese are so popular in Africa, she tells us, because they don’t shy away from infrastructure.

Thank you Ethan Zuckerman.  I wish I could have been there. 
And, I'll definitely follow the future career of Ngozi Okonjo-Iweala.

Build it and they will come - Railroads in the 19th century

While I was in Asia my wife picked up an 1860 original leather bound copy of "The Story of the Life of George Stephenson" by Samuel Smiles.  If you've never heard of George Stephenson, he's known for The Stockton and Darlington Railway which opened on the 27th of September 1825, The Liverpool and Manchester Railway which opened on the 15th of September 1830 and, with his son Robert Stephenson, for the engine "The Rocket" which won the first major locomotive contest in October 1830, also setting a speed record (29 mile per hour).

The Stockton and Darlington was designed to carry coal to a port at Stockton-on-Tees.  To quote Smiles,

At first passengers were not thought of; and it was only while the works were in progress that the starting of a passenger coach was seriously contemplated.  The number of persons traveling between the two towns was small and it was not known whether these would risk their persons upon an iron road. <...>

No sooner did the coal and merchandise trains begin to run regularly upon the line than new business relations sprang up between Stockton and Darlington and there were many more persons who found occasion to travel between the two towns — merchandise and mineral traffic invariably stimulating, if not calling into existence, an entirely new traffic in passengers.

In 1829 the line was extended to new docks five miles down the river where a new town (Middlesborough) grew from nothing to more than 6000 residents within ten years.

For the much more ambitious Liverpool and Manchester, the promoters based their calculations on heavy merchandise traffic (coal, cotton and timber) but also thought that they might be able to capture half of the 400 people per day carried by coaches between the two cities.  Again quoting Smiles,

But the railway was scarcely open before it carried on an average about 1200 passengers per day.

These were heady results that lead in due course to waves of investment (peaking in 1836-37 and in 1846) producing bubbles not unlike our Internet bubble. But despite investment bubbles, and busts, traffic continued to grow throughout the 19th century, literally transforming society. The parallels with the Internet are interesting.  Andrew Odlyzko tells the investments part of the story in section 4 of this article.

What the quotes from Smiles should suggest is we're unlikely to guess what form future advances based on the Internet will take.  Broadband Internet access is important for it's option value, not for the applications we run today.

Put another way, while I love railroads, the Internet's long term benefit for humanity is more significant — likely to rank up there with the introductions of speech, writing and the printing press.

The Black Swan and the Fat Tail

I saw a Guardian Weekly review of The Black Swan by Nassim Nicholas Taleb which mentioned Taleb's reverence of Mandelbrot, thus leading me to purchase and read the book — ok, I read some, I skimmed some.  :)

The central thesis is that wildly unexpected events occur, in finance and elsewhere, more frequently than suggested by a Gaussian or Normal distribution.  In fact markets and many other phenomena are better described by power laws.  Both Gaussian and Power Law distributions have long tails, but the tail of a power law distribution is fatter further out.  That's what makes sales of books and other media on "the long tail" so profitable for Amazon.

Here's a quote from the author:

 The Black Swan is about these unexpected events that end up controlling our lives, the world, the economy, history, everything. Before they happen we consider them close to impossible; after they happen we think that they were predictable and partake of a larger scheme. They are rare, but their impact is monstrous. My main problem is: Why we don't know that these events play such a large role. Why are we blind to them?

It's a book of anecdotes and we've all heard "the plural of anecdote is not data" but while a cute phrase, it's a faulty argument and furthermore there is data for financial markets and long tail markets.  For an accessible version of the mathematics, read the Mandelbrot book I discussed in this post.  But as a book of anecdotes, The Black Swan is a pleasant read and Taleb's primary point is well taken.

90% gross margin on FTTH services

Thanks to Benoît Felten for catching an interesting item in Monday's issue of Le Figaro Economie.  Here's Benoît's summary

< In > an article called Orange, Neuf Cegetel et Free cherchent de nouveaux revenus in Le Figaro Economie dated May 7th, 2007, new Free General Manager Maxime Lombardini indicates that FTTH customers will generate a 90% gross margin. This is to be compared with Free's current broadband ARPU of 33 EUR on which they generate a 20 EUR margin (approximately 61%) on fully unbundled customers... < i.e. ADSL customers >.

Coming from an equipment business with 60%-66% gross margins, 90% sounds rather nice to me.  Of course, that margin has to pay for the cost of capital to build the FTTH network, but at least I now understand why it's reasonable to simplify FTTH return-on-capital calculations.  Assuming all revenues are available to pay off the capital costs is not too far from the truth.

Your mobile is your wallet in Japan

I'm at CTIA.  It's an excellent place to hold meetings with anyone in the mobile industry in North America (and  many from Latin America), but I keep hearing people talk about how cutting edge everything is and I have to wonder at their insularity.  Don't they know anything about mobile phones in Europe, or better, in Asia?  So the most interesting thing in this morning's flood of info was not CTIA related, but this, from Eurotechnology Japan K. K.

The last time I was in Japan, I noticed people wave their mobile phone near the subway turnstile to enter.  That looked like a technology I would really like to have at hand. Indeed, it's not just for subway tickets, but also serves the functions of credit cards, house keys, and company access control cards.  Very, very convenient!

I'm not alone in finding this incredibly attractive.  Adoption in Japan has been extremely rapid, heading towards 30 million by the end of 2007 and a projected 50 million by 2010 (in a population of 127 million).

Now when will we see this in the US?

VON: Cook's BOF - small, but fascinating

Corrected Javad Boroumand 27 March 2007 - rbt

Wednesday afternoon, in the 'VON Theatre' on the show floor, Gordon Cook ran a session entitled Special Cook Report BOF: What does it mean to be an Internet Company? with Javad Boroumand from Cisco, Roxanne Googin, Publisher, High Tech Observer and John Waclawsky, Chief Software Architect, Motorola.

It was relatively small, in a noisy location and didn't answer the question posed, but it was the most interesting session I attended, in part because of a 20 minute discussion with Roxanne and Gordon afterwards.

Javad spoke first explaining what National LambdaRail is, and CAIDA's Commons Project.  The key difference between National LambdaRail (NLR) and Internet2 is that NLR owns their own dark fiber whereas Internet2 is paying for backbone services. This gives NLR enormous flexibility in trying new technology at layers one and two, i.e. below IP.  The Commons project is an attempt to combine the NLR backbone with community and municipal networks that are willing to participate in CAIDA measurement programs, thus getting low level data on Internet performance that supports basic research and lower costs for muni networks.

John went second, speaking about the evolution of edge devices in an open network.  I've heard John speak on this subject and I've talked to him in the past (most recently at F2C), so this was interesting, but not new to me.

Last up, before open discussion, was Roxanne with the financial picture, in part based on her 'Paradox of the Best Network,' i.e. competitive IP networks are not a good investment (capital repellent, in her words) but monopolies (or duopolies) won't give us the broadband access we want and need for economic growth. Unfortunately, she ended by posing the problem but not suggesting a solution or giving any call to action - thus my comments during the BOF and in our discussion afterward.

Responding to questions, Roxanne made it clear she was advocating unfetter municipal experimentation by any of the 22,000 municipalities in the US and/or interested community groups.  I couldn't agree more.  Yes, there will be a lot of mistakes, but to the extent we foster widespread experimentation, workable models will emerge.  Unfortunately, municipalities in many states have been slowed or completely blocked by state legislation and/or state and national regulation, mostly the result of highly effective lobbying efforts by incumbent operators (ILEC and CableCos).

I was struck by the combination of NLR purchasing their own dark fiber and Roxanne's advocacy for municipal experimentation, but I didn't get to push my thoughts until our post session discussion.  My points:

• It takes decades to make fundamental changes in laws or regulatory regimes.
• The right-of-way in front of my home or business is the fundamental access bottleneck.  Conduits and poles in that right-of-way have useful lives of 30-100 years.  Dark fiber has a useful life of 20-30 years or more.  All these align with legal and regulatory timeframes.
• Anything electronic (to light that fiber) is functionally obsolete within 2-3 years.

This suggests the best focus for municipal activities is point-to-point (home run) dark fiber from each business and residence to a central aggregation point where enough other fibers come together that  multiple competitive ISPs (and other service providers) are attracted. Then individuals get to pick which ISP they want to light their fiber.

Of course while my idea of what a municipality should do is right :-), it would be a terrible idea to impose it, nationally or otherwise.  Roxanne is completely correct that we need to give our 22K municipalities permission to do whatever they want and then see what emerges.

Awesome ways to display data

I looked at Gapminder a few months ago, after a mention by Tim O'Reilly (whose post also points to the excellent Worldmapper). 

But today, through a recent blog post by the data geeks at Swivel, I got a pointer to a video of Hans Rosling presenting his awesome visualizations at the February 2006 TED conference.  (It's 20 minutes, but well worth the time!).

Apparently after Hans Rosling's presentation at TED, Gapminder was invited to do a longer presentation at Google.  Then Google stepped up to host the Gapminder site (and presumably to sponsor their work).

By the way, if you are interested in data but not familiar with Swivel, check them out.  I'm looking forward to Swivel becoming the Flickr and YouTube of data.  All they need is some of Hans Rosling's technology.  :-)

Telecom Wrapup — Key Events of 2006

As pundits are thinking about their predictions for 2007, I thought it might be useful to look back on 2006, focusing on items that may not have gotten the coverage — or the perspective — they deserve.

My picks:

1. Death of traditional fixed-line telephony - a turning point
   
2. Dual mode (mobile & WiFi) phones off to a very slow start
3. Some positive results from the U.S. AWS spectrum auctions
4. Questioning the value of Calling Party Pays (CPP) regulation
5. China’s 3G licensing delay obscures interesting VAS developments
6. AdvancedTCA finally gets production momentum

The complete explanations are in my article just published in the NMS Telecom Innovators News.

Foreign Investment in China

It is now well established that investments in telecom are the most productive of any investments a developing economy can attract — better at increasing per-capita GDP than investments in roads, electricity or even education.  So I'm always interested in better understanding what a developing country can do to attrack investment in telecom.  The answer is complex, but a few things are clear.  Regarding telecom directly:  increase competition, reduce restrictions on foreign direct investment (FDI) and reduce taxes on telecom, including less direct costs like spectrum fees and service activation taxes.  But also an issue is a business climate that recognizes property rights and minimizes corruption.  Just how important is that?

I've seen good data to support the benefits of competition, the good results of reducing restrictions on foreign investment (i.e. on allowing foreign ownership of telecom companies) and the positive gains from reducing telecom taxes.  I have been less confident about the issue of good government as China appears to attract enormous flows of FDI despite its government getting consistently bad marks in the western press.  What's actually going on?

My thanks to Ajay Shah who pointed out:

The World Bank has produced a draft volume Dancing with Giants: China, India and the global economy, edited by L. Alan Winters and Shahid Yusuf. Six areas are covered: global industrial geography, competing with giants, international financial integration, energy and emissions, regional variations in growth within the giants, and governance.

Groundwork towards this was done in the form of 21 background papers. Drafts of the book, and the 21 background papers, have been placed on the web. It is good work.

In particular, the working paper Does ‘Good Government’ Draw Foreign Capital? Explaining China’s Exceptional FDI Inflow, by Joseph P.H. Fan, Randall Morck, Lixin Colin Xu, and Bernard Yeung, shows that FDI in China correlates extremely well with respect for property rights.

Their conclusion in response to the question:  Is too much FDI flowing into China?

Probably not.

We have shown that, within China, inward FDI flows disproportionately into provinces with less corrupt governments and governments that better protect private property rights.  We estimate a cross-country FDI model, without China, that explains inward FDI using measures of the strength of constraints on executive power as well as more general measures of government quality and track record in fostering growth and a set of standard controls.  This model predicts FDI flow into China with prediction errors similar to those for other countries with similar levels of institutional development...

There is indeed an enormous FDI flow into China, but given China’s size and growth track record, this flow is not far from what would be expected for a country at China’s level of institutional development. 

So much as we may dislike government repression, it appears the critical issue for raising people's standard of living is protection for property rights and minimizing government corruption.

In any event, there's a wealth of economic information in these working papers.  I'm grateful to Ajay for pointing them out.

More on layer violations in wireless

In late August I wrote on the technology-driven transition from vertical integration to modular industry value chains and how it applied to telephony in general.  A subsequent post on "layers violations" attempted to relate the concepts to mobile VoIP.

I have additional supporting links.  First, my August post alluded to work by Henning Schulzrinne with Andrea Forte and Sangho Shin at the 65th IETF meeting in March.  At that time, the only reference the abstract of a presentation at WitMeMo’06.  The full paper is now on the web.  It's 802.11 in the Large: Observations at an IETF meeting by Andrea Forte, Sangho Shin and Henning Schulzrinne.  As discussed in the previous post, this work points out the extent of problems with current 802.11 implementations, i.e. problems in today's products.

The question for mobile VoIP is whether fast handoff can be achieved at layer 3 (the IP layer) without a tight integration with layer 2 (the radio link).  Traditional mobile networks (GPRS, CDMA 2000 & UMTS) use vertical integration – what a network engineer would call a layers violation – in order to achieve high performance handoffs.  Obviously, WiFi networks (and other general purpose wireless networks) don't do that today.  We could just wait until wireless technology becomes much better – for example until dual radios cost no more than a single radio thus allowing IP session setup over a new radio link while existing IP traffic flows over the existing radio link.  But it could be years before that approach becomes low cost.

Or we could attempt to modularize the "layers violations." The latter is what the IRTF (Internet Research Task Force - the research arm of the IETF) MobOpts Research Group is discussing.

As background there already are Mobile IP protocols defined for both IPv4 and IPv6, but these don't provide handoffs that are fast enough for seamless voice telephony.  Quite a bit of work has been done on IP protocols for fast handoff of IP sessions.  For example, RFC 4068 Fast Handovers for Mobile IPv6, describes what to do at the IP layer,

without depending on specific link-layer features while allowing link-specific customizations.

Unfortunately, to achieve seamless voice telephony handoffs, you will require either two radios or some link-specific customization.  This is where the MobOpts Reseach Group (RG) proposal comes in.

MobOpts RG is discussing ways to standardize the needed link-specific customization – see IRTF draft Unified L2 Abstractions for L3-Driven Fast Handover.  In effect they propose standardizing the layers violation.  If adopted by radio link equipment vendors, this would modularized the value chain without waiting for extra technology to show up!

Of course this is a proposal, not an RFC much less a standard.  So any implementation is at least a few years off.  That's consistent with my view, expressed in my August 31st post, that widespread mobile VoIP will lag fixed line VoIP by at least five years, but it's coming!

Money in dumb pipes?

In the preceding two posts, I've applied Christensen's view, of how technology based products and services are disrupted, to fixed and mobile telephony.  While 3rd party VoIP is not yet dominant in fixed telephony, it's made enormous progress and most people can see the writing on the wall.  But what happens to the lower layers as the value chain modularizes?

In the computer industry, we've seen all sorts of specialization.  But we've also seen companies, like Dell, focus on operations and make money in a layer that R&D-centric businesses avoid like the plague. 

Back in February, Craig Moffett and Amelia Wong of Bernstein Research wrote an interesting paper, The Dumb Pipe Paradox which was distributed to Bernstein Research clients (but unfortunately doesn't appear to be available on the web).  In March, Craig testified on this subject before the Senate Subcommittee on Communications.  At the time, the press reports I saw didn't adequately reflect what was in the original report, for example this from the San Francisco Chronicle:

... financial analyst Craig Moffett of Bernstein Research articulated the dilemma phone companies face as they plan these upgrades.

More recently this from Bill St. Arnaud prompted me to write to Craig Moffett and get a copy of the original report.  It's great stuff !!!

Quoting from Craig's original report:

Our conclusions regarding the economics of “dumb pipe” status are, we suspect, quite counter-intuitive (and starkly counter-consensus).  In short, the economics of such a scenario, as measured by purely financial metrics, are actually better than those of the business today.

• Revenues would fall significantly
• Operating expenses would fall even more significantly
• Operating cash flow would therefore fall more modestly, while operating cash flow margins would  actually expand
• Capital expenditures would fall sharply
• Free Cash Flow would actually increase
• Capital employed in the business would also fall sharply, yielding much higher ROI and ROE

To be sure, our analysis ignores potentially painful milestones on the way to such a “raw connectivity”  future, such as the stranding and subsequent write downs of a huge portion of Comcast’s current asset  base.  Growth prospects in this “end game” would likely also be lower than today’s (although not necessarily).  Nevertheless, our analysis leads to the surprising conclusion that a disintermediated business would, on most financial metrics, at least, actually be a better business than the current one. 

Behind these counter-intuitive findings is a simple, and perhaps more intuitive, truth.  In a high fixed cost business, it’s not such a bad thing to be the low cost provider.

Wow!  Dell's operations focus allows them to make money as the low cost provider of what others consider a commodity.  Similarly in telecom, there is money to be made being the low cost provider of dumb pipes.

Note:  The report discusses both Telcos and CableCos and concludes CableCos would win in the pipes business. But that is very much an artifact of the US regulatory environment.  Think what might be achieved if we had competitive access to first mile rights-of-way.

Layers violations still useful for wireless VoIP

A few days ago I wrote about the observations, by Christensen et al., on vertical integration – when it's valuable and when it becomes a disadvantage.  Their point:  when your task presses the limits of the available technology, optimization is essential and vertical integration is best, as the parts that must work together, i.e. be jointly optimized, are better controlled under the same roof.

As technology improves, you can afford the overhead and obtain the benefits of modularity.  People working on separate modules don't need to coordinate details within their modules, separate companies can specialize, and really big systems (like the Internet) can emerge.

VoIP, i.e. telephony which is separate from the underlying transport, emerged once there was excess capacity in the underlying transport.  VoIP's first major success was in international wholesale (iBasis and ITXC – later bought by Teleglobe) where the VoIP carriers connected directly to the Internet backbone thus getting adequate capacity.  The second major success was IP-PBXs, running on enterprise networks that were over-provisioned and/or had simple QoS (typically 2 levels; gold bits & brown bits).  The most recent success is the emerging consumer VoIP  – Vonage, AT&T CallVantage and the like, and Skype!  These services had to wait for deployment of residential broadband.

Compare this with mobile phone services, mobile Internet and other forms of wireless Internet.  How close are we?  Unfortunately we have a way to go.

Yes, I've used Skype over wireless and, yes, the much hyped IP Multimedia Subsystem (IMS) is based on IP and SIP.  But it's not so simple.

WiFi networks have plenty of capacity if you are sitting in one place with only a few others sharing the same access point.  The moment you roam to another access point you face a gap in connectivity of perhaps several seconds, at least if it's a layer 3 handoff (to an access point with a different SSID) requiring a DHCP request (for a new IP address).  High traffic can cause even stationary web surfing to suffer.  Henning Schulzrinne recently presented work done with Andrea Forte and Sangho Shin at the 65th IETF meeting in March.  They measured hundreds of clients connecting with multiple 802.11 access points.  Among other things, as congestion went up the rate of handoffs soared – less than one minute between handoffs for 22.8% of the clients and between 1 & 5 minutes for another 34%.  The clients are continuously seeking stronger signals and end up thrashing between access points or even disconnecting and then reconnecting to the same access point.  The good news is that most of the problems they found can be fixed with better algorithms using existing 802.11 protocols and some clients, e.g. Apple's, already perform quite well.  But still it will take years to see new algorithms that are also widely deployed, i.e. a new generation of equipment.

With 3G mobile networks, we have different issues.  Yes, IMS is "VoIP", but it's highly optimized and tightly controlled (by the mobile operator).  Thus, neither Skype nor Vonage will use IMS without first partnering (i.e. paying) the mobile operator.

What about straight Internet access?  The good news is 3G networks have been designed to support Internet access, at least asymmetric access for web surfing, and have been optimized for handoff of active IP sessions.  Implementing handoffs has required protocol optimizations (network engineers call these layer violations) and throughput has been optimized via centralized management of client power and uplink scheduling (violating the end-to-end principal?).  Luckily, these optimizations happen below layer 3 – there really is a modular boundary at the IP layer.

The immediate problems are too little capacity on the uplink, at least during busy periods, long latencies and walled garden policies (whether implemented with IMS or otherwise).  With first generation W-CDMA equipment, users share 384 kbps (peak) on the uplink.  That's fine at 2am, but less so when there are others on the air.  Ubiquitous deployment of HSUPA will improve latency and provide adequate uplink capacity, e.g. enough headroom so explicit QoS is no longer an issue for 3rd party VoIP.  I'm guessing HSUPA hype in 2007, urban deployments in 2008 and widespread availability in 2010 and beyond.

Walled gardens are a big issue today, but one that will fall to competition.  Once there is widespread, broadband wireless Internet connectivity from competing mobile operators, WiFi hotspots, WiMAX and other approaches, it will be impossible to sustain a walled garden.  Fixed-line Internet connectivity gives people access to myriad niche services that can't be offered under the walled garden business model.

Widespread use on 3rd-party VoIP on the mobile Internet may lag it's use on the fixed Internet by 5-10 years, but it's coming – IMS or not.

Taxes, bureaucracy and corruption

Catching up on somethings I printed (in June no less), I read this study by the Telecom Regulatory Authority of India comparing the finances of the telecom industries in China and India. 

There a lot of interesting data, but some of their conclusions are amazing.  For example, in comparing mobile operators' OPEX per subscriber, listed as China $4.73 and India $5.49, they offer just one potential explanation

Economies of scale ... could be one of the reasons for lower OPEX for Chinese companies.

Funny, my first thoughts were:

• India's substantially higher taxes on telephony
• India's extra bureaucracy - In ease of doing business China ranks 91 and India ranks 116.
• India's extra corruption, ranking 88 to China's 78.

China might have economies of scale, but that's the least likely explanation, as China Mobile and China Unicom operate with highly regional decision making.

Optimism

This comment by Brian Thomas in Tom Evslin's Fractals of Change is a nice summary of my view of the Internet:

My personal belief here is that given equal exposure, the good generally wins over the bad, and better communication makes it easier for the exposure to be equal, so even though the Internet enables many new ways to degrade, defraud, and deceive people, it also provides new ways to expose evils.

There are many historical examples of this, from the invention of the printing press, then movable type, then telegraphy, telephony, radio, and finally the Internet. Gutenberg did more to break the power of medieval craft guilds and religious authorities, both of whom wanted to keep their knowledge secret.

Power law vs log-normal with really big bucks implications

Yesterday's post was rather rushed as I was about to get on an airplane.  Once on board, it occurred to me, there's another book I've read where the difference between log-normal and power law distributions is key.  That's The Misbehavior of Markets by Benoit Mandelbrot & Richard Hudson.  They argue modeling markets with normal distributions (in price changes, for example) is erroneous and leads economists to underestimate the likelihood of extreme events (like Black Monday in October 1987).

Put another way, small price changes occur very frequently and large price changes very infrequently.  If you plot these on a log-log scale you get a straight line out some distance. Conventional finance assumes this fits a log-normal curve which eventually falls off.  Mandelbrot, argues historical price data fits a fractal or scale-free model, i.e. follows a power law, meaning the likelihood of extreme events has been consistently underestimated.

Those of you in high tech in the US may have followed the accounting change which requires stock options be expensed by the company issuing them.  To do this one must be able to estimate the present value of the options are the time they are granted.  This is done using the Black–Scholes formula.  While the authors received the 1997 Nobel prize in economics (Scholes & Merton; Black had died in 1995) and the Black-Scholes formula is now written into accounting practice, Mandelbrot argues it consistently underestimates infrequent cases of extreme price volatility. 

As an interesting aside, Long Term Capital Management (LTCM), a hedge fund with Scholes & Merton on it's board, collapsed in 1998 due to a series of correlated repricing events that were extremely unlikely, but did occur in the late summer 1998.  LTCM lost $4.6B.

Power laws and log normal distributions appear in many disciplines, not just networking and markets.  In many cases it's hard to determine which applies.  Whether Mandelbrot is correct about power laws and finance, at least we can agree -- $4.6B is big bucks!

Role of language in outsourcing choices

Shingo Murakami, Roger Premo, Ina Trantcheva and Erik Yeager – students at MIT's Sloan School of Management – have just published the results of a survey they recently completed.  They surveyed 60 companies with annual sales between $100M & $1B on their outsourcing activities.  I checked with Erik this morning – these were all US companies.  As expected, language was mentioned as an important issue in forming outsourcing relationships, ranking third after "Specific Technical Expertise" and "Price."  What surprised me was the extent to which language actually drove the selection of outsourcing locations.  Here's the relevant picture:

Note the high scores for the US and India.  It turns out the question in the survey only allowed for the countries listed, but a number of respondents specifically mentioned Canada and Ireland as other important outsourcing locations.

I suppose I shouldn't be surprised.  At NMS we have our own staff all over the world, including major development centers in Montreal and Bangalore, and technical staff in France, Italy, Germany, Hong Kong, Beijing, Seoul, Tokyo, etc.  But everyone we hirer is an English speaker.  We also do some outsourcing to both English speaking nations and to Russia, but again the people we contract with speak English.

A few months ago I spoke with a friend running a high tech company in Germany.  Since the expansion of the EU, he's been hiring engineers from eastern Europe, but only those that can speak German.

I've never found culture differences to be a problem.  Motivated people can easily bridge cultural gaps.  But language is indeed an obstacle, apparently more so than I realized.