Category: Uncategorized

 

IBM, a company now known for computers but once in the more pedestrian business of time-clocks and scales, turns one hundred years old today.  If you consider this a moment you’ll see that makes IBM perhaps the longest-standing tech success in all of history.  Considering the tumult that it’s undergone (you don’t switch from scales to supercomputers without generating some stress) it’s success is even more remarkable.  In a time when many tech companies seem to be floundering, it’s worth a moment’s consideration of how IBM managed to do what it did.  I’ve been involved with IBM in some way since 1964, and I’ve seen almost half its corporate life and much more than half of its life as a computer giant, so let me offer some perspectives.  They might help others who’d like to see their own hundredth birthday.

First, IBM was never hidebound.  Early in its computer age, IBM released a mainframe based on what was at the time the state of the art.  Within a year a revolutionary new option was discovered.  So did IBM feed off the revenues of its just-released product as long as it could and then bring out the new?  No, it brought out the new immediately and wrote off all it had invested in that older model.

Second, IBM was always intensely aware of the buyer’s side of the value proposition.  At every step in their sales and marketing, they supported the decision process from building the business case to installing the technology.  This total-value-chain marketing meant that IBM had a truly unique multi-layered engagement model with the customer, a model that they’ve sustained for fifty years or more.  A model that I’d argue no one else has ever successfully copied.

Finally, IBM has always known the value of and the need for ECOSYSTEMIC technology.  Computers and networks and appliances and other tech elements are not isolated boxes, they’re components of something.  A system of devices needs something to systematize it.

Look in contrast at IBM’s competitors.  Sperry Rand, then Unisys, isn’t a computer vendor any more.  NCR isn’t either.  Digital Equipment Corporation was bought by a PC company, Compaq, who was then bought by HP, the only one of the early players still sharing the computing stage with IBM.  But everything that IBM has done well, HP has done less well.  They have far weaker strategic relationships with their customers, they have made what seems an endless series of critical blunders in tech evolution—WebOS might be the latest—and they’ve mistaken the adoption of a catchy slogan or two as the creation of an architecture and an ecosystem.

You could argue that Cisco, in the network space, is now at a crossroads, a point where it decides whether to be HP or IBM.  Both HP and Cisco have expanded their portfolios without having a strong ecosystemic tie to link their broader line with broader value.  Both HP and Cisco have hunkered down on low-margin products that are becoming lower-margin all the time, just because they’re products they’ve become known for.  Both HP and Cisco had charismatic management that everyone knew, but that built sales and company bulk without creating a longer-term model for success.  So which road will you take, Cisco?  Today would be a nice symbolic day to make that choice.

 

 

There was probably joy in finance-land when Ericsson made their move to acquire Telcordia.  The company, formerly Bell Communications Research or Bellcore, was at the same time the “labs” of the RBOCs and the foundation for the support of and evolution of the classic vision of OSS/BSS.  For years it’s struggled with its conservative roots in a market where old visions and classic visions were increasingly irrelevant.  It was picked up by a private equity player who clearly saw OSS/BSS as more exciting than most of us do, and these guys are probably burning incense right now to acknowledge what they could well be seeing as divine intervention in getting them out of Telcordia alive.  They should be.

So does this mean Ericsson was dumb, and that this is simply a step on the slide of OSS/BSS into strategic irrelevancy?  The answer to both these questions is “Maybe, but then again, maybe not!”

It’s true that there is little in all of networking as boring as an OSS/BSS.  It’s true that OSS/BSS gives glaciers a run for their money in terms of inertia.  It’s true that the OSS/BSS beat is where reporters go if they don’t believe in hell.  But it’s also true that OSS/BSS is the business heart of service providers worldwide.  It’s like demand deposit accounting for banks; you have to have it or you’re not a bank any more.  So while the evolution of the OSS/BSS process has been agonizing, it’s probably still going to happen.  Which is what Ericsson MIGHT see here.

Ericsson more than any other “network vendor” doesn’t want to be a network vendor, they want to be a professional services company.  The task of upgrading back office telecom processes is certainly a major professional services opportunity, and Telcordia is certainly a repository of expertise in that area, a repository with a built-in fan base among the OSS/BSS types at least.  Furthermore, it’s a lot easier to spin a profitable professional services deal for OSS/BSS migration if you have the software components of the new systems.  Any product is good because it can be sold multiple times; services are always one-off.  By blending the two Ericsson can make good deals on OSS/BSS migration for the customer and still have a lot to carry to the bank.

So if this is a good deal, then how will it impact the market dynamic?  Well, you can divide the Ericsson competitors into two classes, those who can only spell “OSS/BSS” and those who actually have a practice in that area.  In the former category you have Cisco and Juniper, and in the latter you have Alcatel-Lucent and NSN.  The move impacts these groups differently.

For Alcatel-Lucent and NSN, this spells very significant competition in back-office integration services.  Alcatel-Lucent, who has been more services-oriented in this space all along, may now find itself looking for some M&A in the OSS/BSS space, and right now they really don’t need the distraction of absorbing something.  NSN has decent OSS/BSS elements and credentials, so it’s risk is competition—not so much from Ericsson (who may be the only company in the space that NSN can stand against in marketing excitement) but because the new market dynamic may require positioning and excitement, something NSN has had problems with from the first.  And of course with Nokia rumored to be wanting to sell off its NSN stake, this isn’t a good time for singing.

For Cisco and Juniper the problem is that neither company wants to think about OSS/BSS at all.  Cisco has always considered it a bastion of the Evil Empire of Telephony, an interesting factoid given that Cisco more than any other company might have been the genesis of the TMF, the Church of OSS/BSS.  Juniper flirted with the evolution of management in its sponsorship of the IPsphere Forum, and that was finally absorbed into the TMF, but Juniper has been steadily disengaging from all that activity and has no relevant position in the space at all.  Both companies could now find their occasional blown kiss in the OSS/BSS direction won’t keep operators happy.

 

 

HP is in the throes of another reorg, this one apparently aimed more at the administrative side than the functional side of the business.  The big change, for example, was a replacement of the company’s chief admin officer and chief information officer.  Neither of these positions are in a direct P&L role, obviously.  One executive formerly heading up Enterprise Services is moving to the board, to be replaced by an outsider according to speculation.

You have to wonder whether this is another of those “we’re really great but somehow we’re executing badly” stories that’s reminiscent of Cisco’s tale after its blown quarter.  In this case I wonder too, because HP had experienced very good gains in strategic influence in our spring survey of enterprises.  CEO Apotheker may be purging some of Hurd’s old hires, but he may also be working to realign the company to execute better in general and eventually to support cloud services more directly, something that would explain the new Enterprise Services exec.

Given that strategic influence is a leading indicator, HP may have some time to do the realignment before deals in the planning phase today actually get to execution.  We noted in the survey that enterprises were rethinking some of their trusted paradigms for data center modernization under the pressure of new insights into the cloud and even virtualization.  That the enterprises expect HP to solve these problems, and even believe they CAN solve them, is consistent with HP’s role as a kingpin IT player.  Since HP is also an increasingly credible provider of enterprise networking gear, it’s also logical to believe that they would be a strong candidate for building a cloud from its constituent parts—a data center and a network.  Lack of actual execution might then be an issue for the future, when enterprises have had a chance to rethink their plans based on their new insights.

That Cisco lost so much ground (to HP and to the field, in effect) in the same survey suggests that its foray into the server space was an initial cause for hope but that the company then failed to deliver what was expected.  Some of the survey comments fit that view; buyers said that Cisco had not offered any compelling strategic vision for the cloud.  Given that HP’s failure in the cloud was a cited source of its current problems, it’s ironic that they had a better cloud credibility rating than Cisco did.

 

Alcatel-Lucent has finally integrated its video story with a partnership with thePlatform to create a broad, multi-screen and multi-service, streaming video system.  The move is likely in response to the rapid development of video project determination we’ve seen in our surveys of network operators.

The “trial marriage” is logical for both parties.  Alcatel-Lucent has a CDN (Velocix) but lacks the streaming ecosystem thePlatform can provide, and thePlatform lacks a CDN and a strong link to the network.  A complete strategy seems critical given the expected trial timing; it’s going to be hard to put together a cohesive story without all the pieces, and an effective linkage between Velocix and thePlatform could really help Alcatel-Lucent with its positioning.  Since virtually every content project we’ve seen is rooted in the CDN simply because that’s where most of the critical components lie, thePlatform gets a big boost too.

One concern I have about the story is its seeming laser focus on the “TV Everywhere” story, which is only a part of the picture.  Most network operators worldwide don’t have multi-channel video offerings today and thus have little or no drive to extend them.  It’s likely that the focus arises from the timing of the announcement with a cable show and the fact that (obviously) cable TV players have a major desire to support multi-screen extensions to their traditional business.

It’s not yet clear just how this whole thing is going to go together, but if we eventually get enough detail on the structure we’ll try to harmonize it with our own benchmark content monetization framework, which we’re publishing for the first time in Netwatcher later this month.

Content monetization is the operators’ top new revenue priority, but two others are also close in their support and actually a bit ahead in terms of timing.  One is mobile/behavioral optimization of services and the other is cloud computing.  Interestingly, all three of these are to a degree converging—or forcing a convergence—in the service layer.  Operators in major market areas are looking at cloud hosting as a framework for creating SaaS partnerships to offer services to consumers, SMBs, and enterprises.  They’re also realizing that multi-screen means mobile screen in most cases, and that supporting mobile LBS and behavior-linked services is likely to require infrastructure that’s more like web hosting than like SDPs and IMS.  Alcatel-Lucent’s content announcement had a little cloud flavor to it and not surprisingly some mobile flavor as well, and this reflects the truth that operators aren’t going to build siloed service layers to support converged networks.

It’s been a week of events and activities, and the only common thread I can find is that they were all pretty much over-hyped.  Apple’s big announcement, the one they trotted Steve Jobs out to keynote, was really nothing more than a locker service with a few tweaks.  Cisco’s event was a capacity upgrade.  IPv6 Day came and went with the world largely unaware, and uncaring.  So is there a message here?  Maybe, and in fact maybe several messages.

Part of the issue here is the need for hype, both on the part of vendors and on the part of the media.  If you’re a tech rag these days (well, “virtual rag” since most aren’t in print any more) you live and die on clicks on stories that can be turned into ad impressions.  Even reporters are often compensated on or at least judged on the clicks their stories get.  Given that, you’d really like people to say something incredibly insightful and interesting so you can start ringing your cash register, and you reward those who do by talking about them.  If you’re a vendor or carrier, you want your stuff in the public eye because that generates leads and also helps grease the skids for sales initiatives.  But if you’re in the PR space you want visibility because you’re judged on it too, just like reporters are.

The problem with this gets us to our second issue.  If the goal is a click, then any click will do.  The old saw about news headlines like “Man Bites Dog” comes to mind.  Why run that story?  Because nobody would think “Dog Bites Man” was unusual and interesting.  But there’s likely more substance to it, so our flight from the pedestrian message shows that substance isn’t what this is about.  That tends to create superficial stories created from superficial messages, simply because superficiality when it’s well-headlined gets more clicks than truth.  And that tends to lead to markets that are under-supplied with value.

Fifteen years ago, the media was the number one source for strategic insight.  Today, it’s the vendors.  In some ways that could be a good thing for sellers; they have the opportunity to create the buyers’ strategic vision and purchase framework.  If they knew how, which is the rub.  The same process that’s shallowed up the press has shallowed up the vendors.

In the June issue of Netwatcher, we’ll publish our survey of enterprise and service provider strategies, priorities, and vendor influences and the results show that in general everyone is losing a bit of influence.  To understand why, we also include a section on how network operators really see content monetization in the projects and surveys we’ve done, and also what’s really happening with data center traffic as we restructure for virtualization and the cloud.  According to the buyers, they’ve communicated all of this to the sellers.  According to the buyers, the sellers are simply not responding, or they’re responding with a suggestion that the buyer simply surrender the whole process to the seller in the form of a fat and profitable professional services contract.  And oh, by the way, nobody really knows what the objectives of the project would be, the milestones or the goal realizations.  Trust us.

And if we want to talk about trust betrayed, let’s look at Net Neutrality.  The FCC, as I’ve often noted, is a Federal Commission and thus effectively a lower court, and its rulings are appealed through the US judiciary.  When the FCC issued its Neutrality Order, it was immediately challenged by players like MetroPCS and Verizon, but the FCC had not yet published its order in the Registry and the courts ruled the challenges were premature.  Well guess what; the FCC STILL hasn’t published the orders!  I don’t remember a single situation in my FCC-watching where an order this important hung fire for so long.  It’s been six months since it was issued, after all.  While the FCC claims it’s not stalling to avoid appeal, I frankly think that’s crap.  That’s exactly what’s happening.  The FCC is desperately trying to figure out a way to make its order pass statutory muster when they know darn well that’s not likely to happen with the document in its current form.  So can they fix it without so much effort and change that it invalidates the vote?  Are they hoping that Congress will act and moot the issue?  There’s little chance that they’d do either, I think, so all of this is just political games.  What a way to deal with a critical public issue!

Maybe they should just create a Neutrality URL to click on.

 

 

We have an interesting counterpoint today in the evolution of technology.  HP’s user conference has been a not-unexpected cloud love-fest, with the company pushing itself as a leading cloud architect and even provider.  And Wall Street is trashing Ciena because they’re not holding onto revenue and profit growth.  At one level, this stuff seems pretty disconnected but it’s not.

HP sells computers and network gear, right?  Cloud computing, if you believe the classic hype, is designed to eliminate private IT in favor of public hosting because it’s cheaper, right?  OK then, how can something be cheaper to buyers without requiring less cost to produce it, especially since the cloud provider would have to earn some profit margin?  Thus, cloud computing in its universal form would create a smaller market for computers.  So is HP presenting a vision of the future where its own revenues fall?

Ciena makes network equipment, primarily deep-layer optical transport.  According to the Street, the company isn’t making enough progress on its next-generation products, products that are designed to offer ever-cheaper bits.  But if video is making the Internet traffic load explode, then why aren’t ever-cheaper bits really important to operators—important enough to be driving up their revenues and profits?  Is it just that somebody else is doing a better job, or is it that even cheap bits are getting too expensive?

What’s unifying these stories is the fact that HP is targeting network operators who want to get into cloud computing.  HP reasons that the build-out created by such a move would drive up their profits at least in the near term.  They also probably believe that the whole cloud-eats-everything story is hype and that it will in fact have little impact on enterprise system sales.  The operators are interested in the cloud because those cheap bits everyone wants them to produce aren’t profitable enough to grow their own revenue and profit and satisfy their shareholders.

But why the sudden push to get network operators into the cloud computing business?  The total potential global revenue stream for cloud computing services according to my model is about $250 billion per year, and that’s about one fourteenth the total service provider revenues.  But it’s at least six times the global online adspend.  What makes cloud computing unique from an operator perspective is that people pay for it; it’s not ad sponsored.  With the sum total of all online adspend far less than the annual capex of operators, they can’t see advertising as replacing pay-for services.  What’s valuable about the cloud is that it taps the direct-payment revenue model, something that online services and OTT players have been largely unable to tap.

We can also see, if we look closely at the HP announcement and also at the announcements of network operators (one of which, Verizon, is a kind of flagship HP cloud account) we can see that operators aren’t trying to make money selling the cloud as much as by selling services of the cloud.  Nearly all the operators who have made cloud announcements are making it clear that their longer-term focus is SaaS, and that they’ll be buying software companies or partnering with them to get services to sell.  That makes sense because if the cloud is a cost-savings strategy, higher-level services like SaaS displace more cost and thus can look more attractive to the buyer.  Enterprises told me in our spring survey that a pure IaaS or virtual-hosting cloud service with enterprise-grade reliability cost them on the average 75% more than it would cost for them to acquire and sustain the services on their own.  In contrast, there are SaaS applications that even for the largest enterprises reduce costs by over 60%.  For SMBs and consumers, some services offer them things they couldn’t possibly produce and sustain internally, for lack of skills.

So what we’re seeing here is another transformation of the notion of what we once called a “service provider” and then a “network operator”.  The shift in terminology came because the OTT players were increasingly the “service providers” and the traditional carriers were simply operating the transport/connection network.  Now they want to be service providers too.

Things are sort-of-happening down at the network level, of course.  Yesterday was “IPv6 day” and like pretty much everything these days it proved to be over-hyped.  The company doing the monitoring for the testing admits that there was only a modest uptick in IPv6 traffic, and most users reported that their ISPs did not apparently offer them IPv6 connectivity for the test (an even greater number probably couldn’t have figured out whether they had it or not).  So is IPv6 another pure hype event?  At one level, yes, but it didn’t have to be.

All the network vendors are obsessed, like HP is, with growing their market.  An enormous network refresh accompanying an explosion in pent-up M2M demand or something, created by the sudden adoption of IPv6, would sure sound good to them.  Next to saying that traffic is exploding, saying that “IPv6 is coming” is a sure path to media attention, or it was.  The problem is that IPv6 is largely a conversion project, and any project manager knows that the best you can hope for in a conversion project is that nobody knows you did it.  The media is hoping for a nice collapse, with hundreds of millions of people crying out in pain.  Nobody wants to oblige, so they’ll stop covering it.

Where the “it didn’t have to be” comes in is that there are potential benefits to IPv6 in security, stability, and performance.  To harness these, we’ll need some changes in the Internet business model and in how ISPs are regulated, to be sure.  We’ll need some of that “put-law-and-order-into-the-online-world” message that we heard from France this month.  But all that stuff just takes sooooooo long to accomplish, and everyone wants instant sales gratification so we won’t go there, and we’ll lose the support of the users of the Internet in the migration of the Internet to the next level.  Doesn’t sound too smart to me.

It seems to me we’re reaching a critical point in sticking our heads in the sand where the whole body sinks out of sight, or maybe where we come through on the other side and see clearly again.

 

Well, Cisco made it’s “big announcement” and it was the second one in two days that fell flat, at least with respect to Wall Street.  It’s my view that it also missed from a market opportunity and requirements perspective.  Cisco moved the ball only a little bit, and it moved the wrong one.

What’s frustrating to me about the announcement, which is a combination of a capacity upgrade to the ASR9000 line and a virtual-device-for-management story, is that it seems to reprise a theme that I’d really resent were I a network operator.  Traffic is exploding, your revenue may not be, but we’d sure like you to support the former trend by buying our boxes and don’t worry about the latter trend.  After all, you’re the guy in the market who’s supposed to supply capacity to meet demand growth.  We’ve shown you demand is growing, so get with it.  Yet Cisco has clear assets to help operators monetize contents, and assets that could be linked to the edge and potentially to this announcement.  They did no such linkage, and thus did a major ballyhoo for a minor advance at best.

All of this came along on the very day a Wall Street report said that Cisco was trapped in an increasingly competitive and commoditizing market.  Well, the ASR9000 stuff sure played into that theme!  More bits, not better bits, is the solution.  But what’s done is done, and now the question is whether Alcatel-Lucent and Juniper will follow Cisco into the hype abyss or be bold and say that the San Jose networking giant is drinking too much of its own Kool-aid.  The challenge, of course, is that for both (but for Juniper in particular) their own Kool-aid is the same flavor.

Virtualization was another news hook for Cisco, the notion of creating a big virtual router to manage, thus reducing the total management tasks to a fraction of those needed to manage boxes individually.  But in truth all virtual-box strategies are little more than embedded approaches to hierarchical management, which goes back to the days of OSI and which is supported by pretty much all Cisco’s competitors in some form.  In any case, the improvement in management complexity is largely proportional not to the number of boxes inside the virtual one but to the ratio between the “exterior” ports and the trunk ports within the virtual box.  A hundred routers inside a virtual envelope is managed as one router to be sure, but managed as one router with the aggregate ports of the hundred.  It’s typically port configuration that’s a management issue, according to operators.

So where does this leave Cisco?  Their assets remain, but I’ve got to be more worried about their will, their leadership.  This, following their Visual Networking Index release, seems to be playing “Chicken Little” with the operators.  If Cisco has the answers they seek (and I believe that the little Cisco may not have could be solved with a wave of their massive checkbook), why rely on thinly disguised scare tactics?  If this is what getting back to basics means, they’d have been better off to stay with adjacencies.  The Street seemed to agree; Cisco shares were off and both Alcatel-Lucent and Juniper were up.

Perhaps they should be up, but as I pointed out earlier in this blog, the response of Alcatel-Lucent and Juniper to the Cisco move isn’t yet out.  UBS pointed out in its note today that Cisco announcements had not led historically to market-share gains, which means that tactical expansion of product lines doesn’t help a company’s sales.  In fact, since the ASR line came out, Alcatel-Lucent has gained share on Cisco and arguably because its own edge products are better-linked to service and monetization initiatives.  Mobile and IPTV success, in short, are driving edge success.  That says that you need a service revenue story to sell boxes these days, and that shouldn’t be a surprise to anyone.

 

Well, Apple had its WWDC and Steve Jobs was there, and for the Apple aficionados it was pure love.  For those less indoctrinated, it was a bit of a yawn.  The iCloud does offer some new things, the most notable of which is an optional system to match on-system music to the cloud’s (better-quality) copy and let users then play the good stuff.  Some are touting this as a way of getting people to pay for pirated material, though twenty-five bucks a year per person won’t exactly stir the heart of the recording industry.  Some think it rewards piracy by giving somebody a good set of songs instead of amateur-ripped copies for a low annual rate.  And functionally, iCloud is still more of a locker; it’s not designed to stream stuff as much as to store it.  While it will make songs available to all a user’s devices, it downloads them on demand rather than streaming them.

The most interesting thing announced wasn’t really even iCloud, it was the addition of iMessage to the new version of iOS.  This will let all Apple device users message each other in encrypted format, with receipts and so forth.  There’s also improved technology to find others that are online, and if anything in the announcement could directly lead to a new service-layer threat, it would likely come from here.

Leaving the song-matching capability aside, iCloud isn’t much different from what Google or Microsoft might offer, and what a host of third-party products also have.  Syncing devices with the cloud isn’t exactly big news; anyone with multiple e-readers does it all the time for books, too.  So we’re left wondering whether Apple trotted out the Big Gun for nothing, or whether the current iCloud is a kind of lightweight shape of heavyweight things to come.

Probably the most interesting stuff was what wasn’t in the announcement.  For example, the regular syncing of devices with iCloud will happen only when you’re WiFi-linked and off the mobile network.  Obviously that relieves what might be considerable user angst over the charges, but it also alleviates operator concerns about the gratuitous traffic.  Operators are also likely to be relieved that video won’t be streamed/synced with Apple TV.  In fact, it appears that Apple may have made a deliberate effort not to push operators too hard.  Might they be waiting until they have something to leverage in such an operator battle?

So adding up the points, you could speculate that the iPhone 5 might be that un-SIM-ed phone we and others have talked about.  You could speculate that when Apple has the tools to cut the cord it will then expand iCloud and take the gloves off.  If so, it would seem likely that the new iPhone and the new independence aren’t too far away.  Why spend Jobs collateral on something that’s not even close to being explosive?  Why alert Microsoft (who sort-of-announced an Xbox TV premium subscription service on the same day) and Google?

The Apple announcement gave both operators and vendors some breathing room; Apple hasn’t made that killer move in the space…for now.  The problem is that the ongoing battle between Apple and Google in the mobile space, and the attempts by Microsoft to elbow in for itself, will surely drive more radical changes, and put more pressure on operators to make the moves their vendors are reluctant to support.  For someone wanting to increase their market share in the router space, this is what to look at.  New models are just new boxes, not new strategies.

 

This is the week of Apple’s WWDC, and everyone is watching to see what Steve Jobs will say about “iCloud”, Apple’s next-gen network service that might be anything from a simple music streaming strategy to an enveloping cloud concept for the consumer.  While iCloud is critical to the future of network services in general and mobile services in particular, it’s not the only issue that’s exploding in the face of the operators.  Disney, for example, is going to start streaming content, and more cable MSOs are lining up to endorse the inevitability of usage pricing for wireline broadband.  We’re at the beginning of what one of our survey operators (quoting Churchill) said was “the end of the beginning” of the golden age of online content.

End?  Aren’t we just getting started?  No according to most operators and interestingly to some of the OTT players we’ve talked to.  Pretty much everyone will acknowledge (in private, of course) that streaming media in any form is a play on the artificial pricing plan for consumer broadband.  It makes no sense to charge zero for incremental usage of something that’s not incrementally free.

Players like Netflix and Hulu and Amazon (and likely soon Apple) are showing us that consumers will pay a hundred bucks a year or so to be able to stream video when nothing’s on TV or when they can’t get to their favorite programs for travel or other reasons.  It’s likely that continued erosion in programming quality will generate some longer-term cut-the-cord trend too, though it’s also clear that you can’t keep streaming old stuff to the same consumers; new content will be required.

The near-term consequences of this is a consolidation game.  Comcast/NBCU is an example of how a company with a lock on distribution branches out to get a position in adjacent market areas.  So’s the interest of access ISPs in CDN services.  CDNs are a natural adjacency for access players, after all.  But in neither of these cases is the move being made today sufficient for the indefinite future, and the players know that.  Apple knows it.

For operators, it’s about getting a piece of this higher-layer pie that matters.  Most operators are now of the view that their role in consumer services beyond voice and chat is very likely to be a wholesale one.  They’re a platform for success, a way of driving service growth by having players with inherently better marginal-ROI tolerance capitalize the knotty parts of the service infrastructure.  If operators can do that, and can do it with the most efficient capital and operations practices in the market, then they can reduce the need for usage pricing and the chilling impact that would have on online services for all.