Month: October 2012

We’re starting yet another trade show in what’s always been the “trade show season” of the fall; this time it’s 4G World.  The timing may be trite in a calendar sense but it’s fitting in a market sense because we’re certainly on the cusp of some major changes in wireless.  The changes may in fact be driving the bigger and broader changes we talk about a lot—cloud and SDN.

One thing that’s clear from the show even now is that we have the same forces at work in wireless broadband as we had in wireline broadband, just wielded by different players.  OTT players were the big push behind the service-model changes of the past.  For wireless it’s the handset/appliance players obviously.  If you look at the commercials for smartphones you see people doing everything but talking, which pretty well sums up the way phones are heading.  The biggest focus is on video, and that of course encourages users to consume it, which makes video the biggest force in disintermediation in wireless, just as it was in wireline.

One effect of this that I can already see in survey results this fall is a considerably greater level of interest in small cells and agile radio networks.  Another is improved integration of Wifi and 4G.  Both are aimed at increasing the total capacity pool available in a given geography so that video traffic doesn’t stall the network.  NSN has been one of the leaders in this space (Alcatel-Lucent the other) and the small-cell trend and NSN’s ability to address it have given NSN the largest improvement in strategic influence of any network vendor except Huawei, and the largest overall for Tier One operators.

Zillions of little cells increase capacity by providing more feed points (backhaul) per unit geography and by improving frequency re-use plans.  They can be tuned to match consumer movement; per-cell using tunable multi-beam antennas and at the macro level by balancing the capacity allocated to different cells.  They also increase the cost of the network, of course.  Still, just a year ago, there were few operators seriously planning for agile RAN strategies and today we have 100% commitment by Tier Ones and nearly that for Tier Two providers.

Another impact of the mobile explosion is a shift in the cloud focus.  I’d noted in past blogs that while operators recognized content, mobile, and cloud in that order as their monetization priorities, they were executing faster on the cloud and slowest in mobile.  There are signs that’s changing now, as operators recognize that mobile broadband services are perfect cloud applications.  A mobile user has an almost transactional need for information if you exclude content delivery; make a request and get an answer.  This model is perfect for a RESTful interface and awful for traditional IMS sessions; too much signaling per unit of service delivery.

Signaling is the other side of the picture, of course.  Not only are services themselves risks to signaling traffic, handsets tend to use more signaling for keep-alive functions—the so-called “dormancy protocol” that’s been evolving as handset vendors try to reduce power consumption during idle periods without losing connectivity.  We think that Alcatel-Lucent’s decision to cloud-host IMS is a critical one because we think that some elastic-cloud vision of IMS is critical in addressing the signaling problems of the network.  Note that this also shows up in the operators’ Network Functions Virtualization priorities.

In the SDN world, mobile is focusing attention on the applications of SDN technology to backhaul and even to mobility, which in our view are related.  One of the challenges of the LTE designer is that the focus of planning has been on “Internet offload” and suddenly applications that look very Internet-like are the revenue source you’re targeting!  We have to think about how mobile content delivery and mobile cloud services are supported—is it EPC or are they offloaded?  If the latter, to what?  If the former, how do we contain costs?  I think that the biggest metro infrastructure question operators have today is how to design an “EPC” that doesn’t present such a cost premium as to make it useless for these cloud-and-content apps.

Clearly vendors will see this particular trend as an opportunity, which it is.  Mobile services already present between 20 and 100 times as many metro endpoints as consumer voice and broadband services in wireline form, and with small cells rushing on I expect you’ll see some areas with a thousand times as many cell connections as central offices.  This kind of change will generate big bucks, and the questions are first how carriers will balance between the layers and second how vendors will present their offerings.  The marriage of SDN and optics could revolutionize metro at the very time when mobile is crying for revolution.

 

The east coast had its “superstorm” and as is often the case people’s view of Sandy varied considerably depending on where they were.  Some areas, including where I happen to live, saw little more than a brisk northeaster, and we were only perhaps 40 miles north of landfall.  Other areas had major damage even though they were further.  You can’t tell about storms.

You can’t tell about markets either.  Apple surprised everyone by announcing the departure of two key executives, one often seen as the most “Jobs-like” of the group.  Some are putting a brave face on the process, saying it will pave the way for the integration of iOS and MacOS into one common platform, but it’s pretty clear that the problem was one of execution.  Despite the fact that Apple is seen as the market juggernaut, the maps issue was definitely a screw-up and Cook seems unwilling to tolerate management that’s too idiosyncratic.  It messes up execution.

Network World offered a view of  what Juniper needs to do to “reignite growth” and it also leads with execution, but while we agree at some level with the points we think most of them are short of the mark at least.  Number one on the list is “execute better” which isn’t much more insightful than saying “sell more”.  Since “market better” is the second point, I guess “execute” means “sell” anyway.  Anyone who ever sold anything knows that marketing and selling depend on POSITIONING, which is establishing a relationship between the product and the buyer’s own value propositions that’s compelling enough to move a deal forward.  That’s the area where Juniper has problems, particularly for its data center and cloud offerings.

NWW’s third point is “sell more QFabric”, which (forgive me) seems like it would fall under the first point.  It is true, however, that you can’t have an enterprise strategy these days unless you have a data center strategy, and QFabric was Juniper’s chance to take the lead in cloud data centers and agile applications, since they announced it before the main spurt of the cloud had happened.  Juniper was actually the first network vendor to talk about what we now call the cloud, at least 18 months ahead of anyone else, but they booted the QFabric launch.  Given that, it’s hard to say how selling more of them could be accomplished.

The fourth point of the piece is “articulate an SDN strategy” and here I agree, but I think the articulation problem is broader.  Juniper has had a problem articulating ANY strategy, largely because strategies are the harmony of products that achieve a buyer goal.  Juniper has been focused on point product from the first, and thus has little interest in harmony of products, and buyer goals are increasingly set higher up on the IT stack than bit-pushing, and it’s there that Juniper has wanted to create value.  The SDN issue is just an example, but I agree it’s a serious one.  While Juniper has talked strategically and at a high level about SDN, at the same time it’s been making nonsensical SDN statements at the product level.

The final point is “better engage with enterprises”, which also seems to overlap the other points.  There are two critical points buried in it, though.  First is that absent strong enterprise sales, Juniper is doomed to decline because it’s clear that service providers alone won’t drive their bottom line to where it needs to be.  Second, “engage” means make a commitment in return for a commitment, and that kind of buyer empathy has been hard to develop when the subliminal goal has been to push boxes on the customer regardless of the value proposition.  What are enterprises trying to do, push bits?  No; Nike and Reebok make sneakers, not networks.  Application software has held up the best in the downturn because application software is about APPLICATION of IT, the foundation of all other value propositions including networking.  The cloud is the closest thing a hardware vendor can get to an application strategy, and SDN is the closest thing that a network vendor can get to a cloud strategy.

I know a lot of Juniper’s sales force; I survey their buyers.  These people are good.  They’re not muffing execution, but they’re being badly served by headquarters, and that’s the thing Juniper has to fix if they hope to reignite their future prospects.  Juniper needs a strategic brainstorm, pure and simple, or no amount of execution is going to help them.

But it’s not fair to beat Juniper up here.  Cisco is sticking its head at least as deeply into the sand as Juniper.  Their cloud and SDN story is better articulated but in a real technical-value sense it’s pathetic compared to what Cisco could do.  Cisco has servers.  Cisco has released a cloud distro.  If there’s a company on earth that should be leading the cloud, it’s Cisco, and yet Cisco is rated below HP in the cloud even though HP has had formidable management challenges and has been perceived as a strategic vacillator.  Alcatel-Lucent has had the assets needed to frame the perfect service provider cloud story for years, and has yet to rise above its own product-silo problems to do that.  Ericsson has the best SDN strategy of any network vendor, and nobody knows about it.  NSN is gaining in buyer credibility faster than anyone but Huawei by riding the mobile/LTE wave, but even though service-layer technology is a small step from where they are it’s a step they haven’t taken.

All of this is demonstrating a lack of buyer engagement, so Juniper has company.  But because Juniper is close to a pure play bit-pusher, they have less latitude to absorb errors, and they are making them as fast as competitors when they need to be doing better just because they lack the mobile and enterprise assets that would cover them for less-than-optimal positioning.  Juniper could tell a great story, if they really tried.  Why don’t they?  Jim Duffy wrote the NWW story, so my challenge to Jim is to find THAT answer and write about it.

The latest story on Juniper/EMC is that there’s not going to be an acquisition but there will be some dancing, primarily aimed at certifying Juniper’s QFabric as an EMC data center switch.  The move is billed by CRN as a “reference architecture” which would imply a much more significant level of integration, but neither the story nor what I hear seems to confirm that.  I’m hearing only that EMC is certifying QFabric as complying with EMC switching requirements.  There’s no reselling, no indication of a broader strategic linkage.

The data center fabric strategy was a lot like many of Juniper’s products—very clever at the product level but launched into a strategic vacuum created by Juniper’s consistent lack of positioning and absence of higher-level value propositions.  There has also been a consistent rumor that the chips that formed the primary foundation of the Juniper launch weren’t even used in the product.  Given that enterprise network opportunity is totally focused on the data center and that the data center is where the cloud starts, Juniper pretty much has to make QFabric work.

Well, I don’t think this EMC thing is an example of how to do it.  I was negative on the launch from the first because of the positioning and the lack of aggressive development of a higher-level context.  EMC might (and that’s a big “might”) add some short-term opportunity to the QFabric mix, but if the problem with the product was strategy and positioning from the first, then only getting those two things is going to fix it.

You have to wonder whether vendors like Juniper resist meaningful cloud and SDN strategies because they don’t understand them, because all this futuristic stuff can interfere with the short-term goal of pushing boxes, or what.  They’re not the only player with the problem either; Alcatel-Lucent has yet to make a meaningful SDN statement though I hear they are working feverishly on it.  Ericsson has made some very smart and timely comments on “service provider SDN” but so far hasn’t provided any meat on the bones of the story (and I’ve asked for it).

This sort of behavior may well be a key factor in the operators’ launching of the Network Functions Virtualization (NFV) initiative.  The opening paragraph of the white paper produced by the group talks about the explosion of proprietary, purpose-built, gadgets in the network and the cost and agility risk they pose.  You have to believe that operators writing a paper would start it with their pet peeves.  And NFV is a considerable step toward commoditizing network equipment, something that SDN by itself could be but wouldn’t necessarily have to be.  The sense of the NFV paper is to move functionality out of proprietary devices to be hosted in servers.  Cisco’s whole SDN direction is about proprietary functionality, and Juniper’s recent Universal Edge announcement is about moving things from servers into proprietary devices.

Of course, Juniper could revolutionize networking by pushing NFV principles themselves.  Why?  Because they took a big step in that direction years ago.  As has been the case so often, Juniper had a product that anticipated the current market trend.  It hosted Junos on what was effectively a separate server.  Where did you go astray, Juniper?  There was so much potential there!  To close out the theme, they went astray by depending on partners for application-layer stuff, and that never moves a company’s strategy forward.  Now they’re doing the same with QFabric.  So when the same CRN article notes that Juniper is undergoing a high-level management restructuring, I have to wonder if the people are getting replaced because they didn’t drive good strategies and positioning, or whether they’re dumped for failing to make the bad strategies work.

The parade of misses continues, so it seems, with Acme Packets, Amazon, Apple, and Ericsson joining in.  One might think this proves that economic conditions (those “macro” factors I blogged about yesterday) were to blame after all.  I still don’t buy it, and if you look at these companies you see a mixture of root causes.  Yes, more sales would have covered them all, but it’s in getting more sales that companies distinguish themselves.

Amazon clearly can’t expect to own all of retail, or even all of online retail, and its moves into tablets demonstrates that it realizes that.  The company is spending more now, including acquiring a share of social-coupon player LivingSocial, and that’s reflecting on the bottom line.  Even without LivingSocial considered, they’d have posted a net loss this quarter.  That’s a signal that Amazon, cloud included, isn’t growing top line at the same rate as costs.  Economic improvement would likely have pushed facing reality out a couple of quarters, but either some of these new ventures start paying off handsomely for Amazon or they’re going to start disappointing regularly.

Apple is in somewhat the same condition.  How many iPad models or iPhone models can you put out when your baseline business model is to sell to a relatively closed community?  The pricing on the iPad Mini proves that Apple isn’t trying to steal share from Google or Amazon, they’re trying to defend their current Apple Basketers.  Yuppies apparently are born and not made and that confines Apple to a limited TAM.  Do they realize this and are simply milking the last buck out of the traditional fan base before cutting prices and foregoing the “coolness” angle, or will they do the same thing in tablets that they’ve done in PCs and phones—surrender the real market to another player?

Ericsson, I think, is a victim of itself pure and simple.  The company has incredible assets in the space, it’s the largest equipment player in networking (for how; Huawei is waiting in the wings), and it is actually doing good work in all the hot market areas like the cloud and SDN.  But the company is so far from self-promotion it’s bordering on being self-effacing.  The best example of this is the fact that a bunch of operators have now stepped up to drive the Network Functions Virtualization model of NGN service architecture, even taking the venture to live inside ETSI.  Ericsson could and should have launched that initiative and made that ETSI proposal.  You don’t have to market to win as a price leader, but you do if you aren’t.  The reduced margins this quarter show they’ve surrendered positioning advantage and are now butting heads with Huawei, a battle they cannot win.  So fight the one you can!  Learn to sing.

Acme Packets has no economic problem at all, they have a major positioning problem.  There is absolutely no hope that voice services can ever be truly profitable again unless you have a cost model that lets you give them away like Skype.  Acme is hoping that LTE and VoLTE is going to ignite their business, but the voice business MUST shift to app-based voice, as I’ve said many times before, because no mobile device vendor is going to be able to keep cheap voice apps off their devices.  Furthermore, if Apple and Google decide to become MVNOs (which I think is a given) then they will exercise OTT voice because that’s their best option in a profit sense.

OK, here’s the straight dope.  Service providers are not growing capex even if they increase profits (Verizon said that explicitly) because ROI on transport/connection infrastructure is falling.  Instead they’re focusing on building service-layer stuff where ROI is higher.  Five years ago, they’d have KILLED for a proprietary strategy in the service layer as an alternative to what they got from their network vendors, which was no strategy at all.  Today they’re going their own way.  Telefonica Digital bought ANOTHER software player this week.

Enterprises have driven their budgets for three years on virtualization, which is first a cost-savings strategy and second a strategy that has a natural end-point.  When you’ve virtualized everything, unless you’re an idiot in project management, you’re now running a lot less servers.  Even normal “refresh” is now going to cost you less.  You can’t EVER grow a market through cost management.  Enterprises can make real benefit cases for the cloud, but we’re pushing more cost management instead.  Cost management vanishes to a point—zero cost.  That means zero spend.  The next time a vendor tells you that their strategy is to manage your costs or TCO, check their passport and if they’re not from China you’re wasting your time.

A recovery won’t recover sanity for vendors who are still mired in the old world.  Even a flush buyer is smarter than a stupid seller.

The earnings reports from F5 and EMC make it clear that tech in general, and networking in particular, and enterprise most particularly, has serious challenges to face.  The largest one, I think, is that people have gotten used to the idea that networking was somewhat immune from macro conditions; clearly it isn’t.  But macro conditions don’t tell the tale in my view.  The biggest problem we have in networking is a case of the Fundamental Stupids.

Every project is a cost-benefits analysis, right?  OK, then, if we want to improve cost-benefit at the project level and raise our company’s rate of return on invested capital, we either have to raise the return or lower the investment, right?  Simple math.  ROI equals R divided by I.

The Fundamental Stupids disease makes seller executives immune to this equation.  They foresee increased I without any gains in R.  They think that if you have infrastructure that costs you a million bucks, you can somehow toss it and put in new stuff because it’s “cheaper”.  The truth is, and has always been, that you can’t improve investment in anything if you can’t increase the benefit case.  The truth is, whatever you already have is nearly always cheaper than something you buy to replace it unless you wait until the end of the depreciation cycle.

That’s the underlying problem these days.  It’s been a full decade now since we’ve presented buyers with any new benefit case for networking.  The IT guys have done better with things like Big Data, but even there the best efforts haven’t really moved the ball much.  This, despite the fact that there are many examples of companies making major gains in productivity by doing something innovative.  In my spring survey, I found that 13% of enterprises reported having improved their tech ROI by 15% or more by increasing productivity.  That means 15% more spending can be justified.

“The cloud” is the current chance to pull the industry out.  No, not by making IaaS the alternative to server consolidation.  Thinking EC2 will transform IT is another example of the Fundamental Stupids.  What transforms IT is new stuff to do with it, valuable stuff.  The cloud, by making IT elastic, makes it possible to empower people when and where they need it without wasting a lot of resources empowering them when they don’t.  By combining the cloud with mobility, you give every worker what I’ve been calling their “jobspace” of information and process empowerment when and how they want.  Even for existing processes that would be helpful, but just as we’ve re-learned communication at the consumer level based on smart mobile devices, we can re-learn how to conduct business.  This is the pie we should be cutting up in earnings seasons, the market we should be addressing and reporting on.  It isn’t, and that’s the fault of the industry and not of the “macro environment”.

Microsoft brings out its Windows 8 and tablet products officially today.  If ever there was a company whose life depends on a radical rethinking of benefits, Microsoft is the one.  Apple and Google have won the traditional appliance market, and Microsoft can’t reverse that unless they want to sell at a loss.  So why not go after the non-traditional market?  If Windows 8 on tablets and PCs can unlock some of that new cloud benefit, then it’s a winner.  If not, no matter how neat the keyboard is or how “convenient” the GUI is, it loses and Microsoft loses along with it.  Where, though, in all of Microsoft’s market hyping, is there a comment on anything “new” except in a technical sense?  Novelty can be either a trinket or an expression of innovation, and what makes the difference isn’t how it’s accomplished but WHAT IT DOES.  Utility is the key to novelty; something that’s useful and novel is revolutionary.  Something that’s just novel is our generation’s pet rocks and hula hoops.

To circle back to networking at my close, I point you to Light Reading’s article on SDN today (http://www.lightreading.com/document.asp?doc_id=226325&site=lreurope&).  The main points are comments like “possibly the biggest shift in telecoms in 30 years” and “this is serious stuff” and “expect things to happen fast”.  The Network Functions Virtualization activity by operators is a badge of shame for network vendors.  For FIVE YEARS those operators have tried to get their vendors off box-pushing and TCO babble, to no avail.  Now they’re moving on their own.  It’s not that operators intend to commoditize network equipment, but that the commoditization is the natural result of a football game if one team refuses to take the field.  Imagine what would have happened this quarter had ONE VENDOR simply done what operators were asking.  Would that vendor have been talking about “challenging macro conditions?”  I doubt it.

The Network Functions Virtualization stuff I talked about in my blog yesterday has picked up steam, adding more Tier One sponsors and turning to ETSI to host them as a body.  The group is emphasizing that they’re not a “standards group”, hopefully trying to avoid the glacial pace and lack of market responsiveness that characterizes standards these days.  Today, standards are to market needs as forensic pathology is to disease; it’s too late to help.

Stu Elby from Verizon offered a presentation that makes a clear distinction between what I’ve been calling NFV and what the market calls SDN.  One of his slides shows services and applications talking to an orchestration process that then talks on the right to the SDN/OpenFlow world and on the left to the cloud/DevOps world.  To me this is the heart of NFV; networking is more than bits, so network infrastructure is more than network equipment and service creation is more than circuit/forwarding control.  We’re actually, via the NFV movement, on a track that just might give us the architecture for the “NGN”, the service network that finally replaces the PSTN and lets the operators play in the new-revenue sandbox.

Stu’s first slide was a price-cost curve that showed the compelling problem driving all of this.  Revenue per bit has been falling steadily for a decade, and while cost per bit has also fallen it’s not falling nearly fast enough.  The operators believe that means it’s time to create a new architecture where boxes (switches, routers, optics, whatever) are plug and play, where IMS is a cloud app (Elby showed Verizon’s progress with that).  Host even IMS on generic servers?  That’s an interesting development, and it shows how badly network vendors have booted this one.  They should have been the ones taking this notion to ETSI, they should have (and could have) grabbed leadership.  Instead they manned the walls while the foundations eroded.

Case in point:  Juniper reported its quarter, and the results missed consensus expectations narrowly.  The results offered a mixed view of the company to financial analysts who crunch numbers, but a pretty clear view to people who focus on fundamentals.  More than any other network vendor, Juniper is vulnerable to commoditization because it lacks any conspicuous position in the money-spaces of the market, notably mobility.  It’s suffered continuous loss of strategic influence in the operator and enterprise spaces in our surveys.  Now the problem is that while Juniper could have done truly great things a year and a half ago, they’re largely caught at this point.

Operators don’t want to host services on edge routers; the whole NVF movement has made that clear.  The PTX is a lower-cost core-network alternative, so wins there are losses in an overall sense.  QFabric wasn’t built as an SDN strategy; there was nothing whatsoever said about SDN at its launch, only about the chip that rumor says isn’t even being used.  Production versions of OpenFlow don’t deliver on SDN, they just deliver on a tiny piece of it.  In fact, they don’t deliver on the SDN vision Juniper has talked about in public before.  Instead we’ve had Cisco-like sponsored research on how traffic is growing, or the now-traditional Juniper assertions that they’re the TCO champs.  Operators agree that traffic is growing, but they’re trying to counter that with LOWER cost, and they don’t believe TCO numbers and neither do I.  And even on the earnings call, Juniper had to acknowledge some leadership changes as executives and lower-level key people depart.

There’s trouble here, pure and simple, and it’s trouble that is totally unnecessary because the company had a compelling story and opportunity just a year and a half ago.  And Juniper isn’t the only one with a problem.  Every vendor is still dragging their feet and watching their customers take over a leading role in defining the network of the future.  Then they’re surprised and horrified when that network is made up of interchangeable commodity boxes and server-hosted features.  Give me a break!

Apple had its own moment, even though it wasn’t an earnings call.  The much-anticipated iPad Mini came out, and the product surprised a lot of people.  The new tablet is bigger than a Kindle Fire or Google Nexus, has a lower screen resolution, and is priced significantly higher.  It’s clear that Apple isn’t going after the mass market in 7-inch tablets at an over three hundred buck price-tag given incumbent products at two-thirds that price or less.  What the heck are they going after?

The “Apple Basketers”.  There are people who will buy an Apple product in a given space no matter how it’s priced relative to the competition.  They can’t do that if there’s no Apple product in the space, though, so Cook is obliging them—with a bunch of new niche-targeted stuff, in fact.  The goal is to tap off Basketers who will otherwise defect to Amazon or Google, which in the long term helps Apple keep an ecosystem intact.  That’s their underlying strategy, and it’s not a bad one.  Why be a market leader in a market where products are discounting to dirt?

But the big miss to me was iCloud.  Apple still has perhaps the most pedestrian vision of the cloud that ever was, something that’s really surprising given the company’s record of innovation.  Well, maybe not too surprising.  Cloud-based services undermine the value of handset and tablet incumbency by focusing the user on the network and not the device.  What Apple did in my view was to tell the world it’s going to be an MVNO, a handset player who starts selling their own mobile services based on carrier relationships worldwide—relationships some people tell me are already being discussed.  Apple won’t make the cloud the star till they own at least the customer relationship at the service level.

 

It seems the SDN scene just never goes to sleep.  Today at the SDN and OpenFlow World Congress a group of Tier One operators announced they were launching an ambitious initiative called Network Functions Virtualization (NFV) to standardize network infrastructure and virtualize network functions/services.  As I’m hearing, this activity goes beyond OpenFlow and SDN to address standardized cloud infrastructures and even a possible PaaS-like structure on which to build services.  However, it’s not clear just how far it will go in this ambitious direction, or how fast.

What seems to be the key point of NVF is the abstraction of network infrastructure functionality, something that could in theory reduce it to a model whose component elements had specific interfaces and functionality.  These elements could then be mapped to equipment with the assurance that anything that represented itself as “Element X” would perform like any other option, and so components of infrastructure could be plug-and-play.  This is clearly something vendors would hate, and just as clearly something operators would love.  I love it too because I’m a big fan of the abstraction/decomposition model in general.

The challenge here is to get the abstraction model going.  “Analysis” in logic is the process of breaking something down to constituent elements; “Synthesis” is the process of creating something by assembly.  We seem to get stuck in synthesis in networking, starting at the bottom and building up to the top.  In community activities, as this one already is, the risk is that you assemble legs and trunks and tails and never quite come up with the elephant.  If you proceed on an analytic basis, you start with the high-level goal, which makes it harder to get off-track.  Vendors typically create the synthesis pressure because they want every story to be a box story.

The thing is, there’s a lot of open real estate at the top right now and a lot of disorder at the bottom.  A vendor who can simply rise above the din and get to the value proposition has the opportunity to create all the important elements of NFV before anyone get there.  You can see a bit of this in how SDN principles fit into transport networks, a topic that was the subject of a Hangout on Google+ with Ping Pan (IETF and OpenFlow guru) and I (http://www.youtube.com/watch?v=5kSnxsYyjiY).  The same video demonstrates the “synthesis/analysis” dichotomy.  We can start talking about how to fit OpenFlow to a transport model, but we can also start by talking about what’s different in transport SDN than in “packet SDN”.  You decide which you like!

Cisco, I think, has aspirations at the higher level of SDN.  In any realistic SDN model, SDN has to start with the cloud.  That means you have to map the virtualization that SDN uses to represent the network’s services to the virtualization that the cloud recognizes.  Right now, the latter is represented in OpenStack’s Quantum interface and in a number of DevOps projects that are dealing with cloud deployment and (of necessity) how network connectivity figures into that.  If you start there, you can visualize a cloud network as a virtual structure with abstract properties, and you can visualize provisioning as an orchestrated process that assigns resources to meet those abstract goals.  NVF seems aimed at generating just this sort of thing, but is it going to do it by starting with the cloud or starting with the network?  If the latter, then Cisco may own all the useful stuff before anyone even realizes it’s needed.  Which, of course, is likely what Cisco wants.

NVF is the product of some of the largest Tier 1 providers in the world, and if they really get focused on the goal line they’ll make a major difference.  Perhaps the most significant difference is that they’re going to level the playing field in terms of the mapping of functionality to equipment, insuring that vendors are tested on common capabilities.  That shifts the notion of differentiation from box-to-box competition (which is possible only on price in an NVF world) to strategic positioning of your products to meet the functional goals NVF provides.  In other words, it’s time to get out of the bit weeds and think about clouds, network vendors!  That’s where the definitions that set your future will come from.

ATIS (Alliance for Telecom Industry Solutions) is going to launch an SDN initiative aimed at generating a viable ecosystemic strategy without the bottlenecks of traditional standardization—take a leaf from the OTT tree.  This is sort of hopeful to me not because I think they’ve made any progress at this point, but because they seem to be acknowledging the two critical data points that other activities have forgotten.

We’re going to have SDN ready or not, standards-wise, because market conditions will drive progress.  Cisco made that point in a negative sense by demonstrating it’s prepared to meet SDN goals on its own, both in the optical space and in virtual networking.  Ericsson and Huawei both took the positive side of the issue, speaking at BBWF in favor of doing something constructive in a collaborative sense in the industry.

I’ve written here often on the need to look ahead for the cloud, to a time when applications are developed with the cloud in mind.  That view hasn’t gotten much (well, any) play with the major cloud types, but now we have a cloud player who’s saying the right thing (sort of) about the cloud value proposition.  Cloudscaling, an experienced player but hardly a household word, actually features the notion of dynamic cloud-specific applications on their website, making them the first of the cloud stack providers to push that critical positioning overtly (others could argue they support the concept but don’t call that out in any conspicuous way.

The Cloudscaling “Open Cloud System 2.0” was announced last week at the OpenStack summit, and that demonstrates that the product is built on OpenStack.  The material suggests that Cloudscaling is adding a management/abstraction layer that is designed to help host SaaS and PaaS apps as well as those dynamic for-the-cloud applications, but there’s no real detail provided on exactly what is offered or how it works.  Given that OpenStack is essentially an IaaS model, it’s not clear whether Cloudscaling doesn’t tout IaaS because it’s OpenStack table stakes or because they have something truly special built in/on it.  I’ve asked the company for more information and I hope to be able to dig a bit deeper when/if they provide it.

Whatever the secret sauce is, the company has managed to gain some traction with network operators.  They are one of two lesser-known cloud software players who actually show up in an operator survey of cloud software (Joyent is the other), and their message that what gets built as applications and services on the cloud is more important than hosting IaaS resonates with operators I’ve talked to, most of whom are very concerned about having cloud computing become just another low-margin business like transport/connection networking.

I think the confusion about just what these guys do is a testament to the problem with the market perception of the cloud.  I’ve talked to a lot of reporters and editors about the cloud, and while some are very open to hearing views that diverge from popular cloud myth if they’re grounded in real buyer/provider interest, others are simply trying to write another chapter in the same story.  This makes it hard for a company like Cloudscaling to get a new story out, even if they have all the collateral in the world to justify it.  It also makes it hard for cloud prospects to review candidates for their infrastructure without going through an arduous research process that involves contacting everyone and asking for answers.  Most early-phase buyers are reluctant to step in front of a sales-call avalanche by making direct contact and asking for data, so in many cases these companies never get to bat.

One company who is getting to bat a LOT more often, apparently, is NSN.  After years of being a drag on the parents’ earnings, NSN has made a nice contribution to them in the current quarter.  The move reflects the company’s focus on 4G technology, and in particular on the use of a combination of smaller LTE cells and WiFi to create a highly flexible and even higher-capacity mobile network.  NSN makes the bold statement that they could deliver a gig per user in quota bandwidth, far more than is consumed today but consistent with what large-scale video use (tablets, for example) could be expected to drive.    It’s a start of a good story, I think.

The other part is something in mobile/behavioral services.  Operators are not at all sure that the current model of mobile marketing (“Who has the latest iPhone or iPad?  Me!”) is going to survive long given that Apple and others who can build their own brand can easily consider MVNO status.  That renders the operator invisible, and that means that talking about how fluid their mobile network is won’t do much good.  Operators need to be thinking about how to make their network assets rear up above the waterline in a handset-MVNO world, and services are the only solution.  If NSN gets that part right they could be formidable.  Maybe they need a cloud partner!

 

Anyone who thought PCs were healthy has probably had that knocked out of them by now.  Microsoft and AMD have added their voices to the chorus of “below seasonality” qualifiers as the companies reported lower numbers than expected.  AMD is clearly in trouble, with significant layoffs now on tap.  Google’s numbers, released by surprise in the middle of the trading day, were below expectations and the search giant took a big hit.  Alcatel-Lucent and Nokia are laying off; Juniper is rumored to be on the block.  IBM’s numbers missed.  “Great Caesar’s Bust is on the shelf, and I don’t feel so well myself”, to quote a whimsical poet.  What gives?  The answer is “mobile/behavioral symbiosis”.

Nobody believes that the growth in PC usage over the 2000s was due to millions deciding to write their memoirs on word processors.  PCs were sold as on-ramps to the Internet, and as many as a third were never used in a significant sense for anything else.  Why?  Because that’s all there was.  When smartphones and tablets came along, they started to cut into the “I want to be online” market, and that’s going to hit the PC space.  PC sales are off.  They are going to stay off.  If I’ve been hammering nails with a wrench because hammers weren’t invented, I’m not going to go back to wrenches in the hammer age just because somebody puts a claw and a head on one end of the handle.

Which is the issue for Microsoft and the PC chip companies.  Ultrabooks won’t compete with tablets, period.  Nothing that you do to Windows will capture a buyer segment that doesn’t want to run any OS at all, they just want to be online.  Even Google with Chromebooks can’t turn back the clock.  A third of the PC opportunity space is gone, most growth will be in that segment, and the PC space will never be the same.

Does that make Windows 8 a failure?  It depends.  If Microsoft thinks W8 is going to recapture the past, forget it.  That never works in any market.  If they think that W8 will leverage their current Windows franchise to create familiarity with a new model of GUI that is portable to phones and tablets, and thus help Microsoft cross over that gap, then they have a shot.  They are going to lose money in the near term as people resist the W8 look, and they risk having a flop like Vista if they’re not careful, but probably they can ride out the transition and emerge stronger in tablets and phones.  Not stronger overall, or even strong in an absolute sense in tablets and phones.  They’ll never be more than number three, but they can hope to be a viable number three.

How about Google?  They have Android, which is number one in phone OSs and at least a contender for tablets.  But they don’t really make money on it, and outside their Motorola unit there’s going to be less on the table for Google and not more.  Amazon’s Kindle Fire demonstrates the problem with Open Source; you can’t sell it and others can use it freely.  It will be very easy for strong players to build “shells” or “GUIs” on top of an Android kernel and run wild and free in some non-Google direction, which is what Amazon is doing.  But it doesn’t matter because for Google it’s search or nothing.

Big news; all online advertising is a zero-sum game at this point.  We’re seeing, in Google’s report, a lower price per click for advertising.  That’s a commodity market, which doesn’t happen when there’s dynamism, growth, and new TAM to address.  Yes, there are still areas where online advertising can grow, but those areas won’t make up for the increased competition overall.  And while search is still the “best” mechanism for online advertising, it’s got two fatal flaws.

Flaw number one is that you don’t search much when you’re mobile (remember, mobile/behavioral symbiosis?)  The mobile user has an instant gratification mindset because they don’t have much ponder time.  That’s why they bought apps by the millions to do what could have been done on the Web without spending a nickel.  Mobile users are not only limited consumers of search, they’re a training-ground for non-search behavior.

Which brings us to Flaw Two.  I do two kinds of searches; one for research, where I get some ad-linked or optimized results that really annoy me, and one for products, where all I get are ad-linked or optimized results that annoy me.  Thus, I’m perpetually annoyed.  How do I escape?  I reason that there are only two online sources likely to give me a good price and reliable service—Amazon and eBay.  If I simply go to my favorite, look up the product, read the reviews, and buy or not, I’ve eliminated all those annoyances and also eliminated my component of Google’s search ad revenue opportunity.

So does Google get into social networking?  They tried, of course, but that’s up there with Microsoft or Intel putting hammer-heads and claws on wrench handles anyway.  There’s no recovering lost market opportunity; you find new stuff instead.  That’s what everyone in the market, including Alcatel-Lucent and Juniper, need to be doing.  We will never go back to the past, but the mobile/behavioral changes in us all are creating a transformation of goal fulfillment that only the cloud can address.  We will have more for-pay services in the future.  We will have more voice activation, more convenience, and all of this simplification of our use of technology has to be made up by making technology more complicated in how it serves us.  That demands centralization; whatever you think about an iPad it’s not going to store the knowledge of the universe (nobody will pay usage rates to update it).  It doesn’t have to, if it can ask for it via super-Siri and the cloud.

The cloud is where “network value” has to go, because the cloud is the new generation of “online” that users continually chased since the dawn of the Internet.  It’s the architecture to create the “Universal Constant”, that number which multiplied by my answer yields the correct answer.  What I need and can’t provide myself, or even ask for in an organized way, the cloud giveth.  Thus, if you’re a network vendor THE ONLY THING THAT MATTERS IS YOUR CLOUD POSITION, and that doesn’t mean you say “I network the cloud” and instantly have one.  SDN is the path to creating a cloud mission for the network, so you don’t have a prayer of getting a cloud position without an SDN position.  Who, of the major network vendors, really can say they have one?  Not just an SDN billboard, an SDN position!  Thus, nobody (even Cisco) has a real cloud position.

Mobile broadband ties our information tools to us, so it ties us to our tools.  It’s changing how we live, and that changes whole markets—markets far removed seemingly from little square devices we call smartphones or tablets.  It changes how we buy, how we eat, how we learn and teach, how we find our way,  make friends.  This is a revolution, and we’re seeing the signs of major incumbents not getting the message.  Some will eventually, and perhaps even lead in the future, but some clearly are already too far gone.  Which are you, dear vendor?  Think about it.

 

The big news today is the rumor on the Street that EMC is about to do a deal to acquire Juniper, a rumor that sent Juniper up about 7% pre-market today.  This has been speculated on before (Light Reading did a piece, for example), and there are some seemingly strange things that such a deal could explain.  One is why Juniper went silent so suddenly on SDN, appearing to stall all its strategic positioning while continuing to make the usual box-pushing announcements.  Another is why the stock seemed so resilient recently given continued speculation that both enterprises and providers were holding back on spending.

The biggest news in this deal would, of course, be the notion that EMC’s VMware could be better positioned to go after Cisco.  Given that Cisco has recently been pulling back from VMware, relationship-wise, this is another interesting data point because it suggests Cisco might believe that VMware is going to be more a problem than an ally in the future.

On the surface, the deal would have possibilities because it could create a network/SDN vision that would link software and hardware in one company.  While EMC doesn’t have servers per se, they do have a decent position in the data center, and in our most recent survey they suddenly jumped in credibility based on their VMware position.  The burning question would be whether EMC could quickly change Juniper’s positioning, which has been far too pedestrian cloud-and-SDN-wise to allow EMC much of an upside.  That would probably depend on what might happen with management and the board.  Sticking Juniper-as-it-is under an EMC/VMware brand would accomplish absolutely nothing other than make Cisco—and likely IBM, and HP—intractable enemies.

If the deal gets done, that is, and in my view this is a VERY BIG “if” indeed, for all the reasons I gave when the notion first surfaced in LR.  EMC is still operating at a P/E far lower than Juniper and so the deal would be rather rich for them.  There’s also an industry aversion to the whole storage-net-buys-regular-network play based on Brocade’s Foundry experience.  Not to mention management challenges, integration challenges, and the fact that the deal would put EMC in the service provider network equipment sector more than in the enterprise sector.  They have no experience in the SP space.  What’s my view?  I think the deal is TOO rich for EMC’s blood.  I think this is just a rumor.

It’s study time, cloud-wise!  There are a couple of new and interesting ones that have come out recently and I want to take a minute to comment on them since my own fall survey of enterprises is now underway.  We’ll be reporting results in our December Netwatcher issue.

The first study is from ISACA (http://www.isaca.org/Knowledge-Center/Research/ResearchDeliverables/Pages/2012-Cloud-Computing-Market-Maturity-Study-Results.aspx) shows a number of interesting things.  One is that the cloud market is not yet mature; in fact, it’s in its infancy (which means that calling Amazon the winner is decidedly premature).  Another is that business enablers are a bigger driver of cloud adoption than “financial considerations” meaning cost.  The biggest business factor is improving availability and quality of experience for applications.

There are some other interesting elements in the ISACA study too.  One is that only about a third of respondents said the cloud market had a high level of innovation, and a fifth said it showed “limited” innovation.  Another is that cloud users were more confident in their strategy than in the service or problem resolution capabilities of the cloud operators.  I think that the study points to an industry that is convinced the cloud will mean a lot for them, but also of the view that it’s not yet delivered on its promise to the market at large.  It’s not that they’re unhappy with the cloud (the numbers on service satisfaction in the study match our figures for internal application satisfaction almost exactly) but that they’re not far enough along yet to be fully convinced or committed.

Given that one of the findings of the ISACA study was that PaaS was more an infant than the other cloud service models, the IBM study (http://www.ibm.com/cai/paas) is interesting in its focus on PaaS.  In fact, the study suggests that PaaS is the logical way for enterprises to migrate their OWN applications to the cloud.

PaaS is underrated, I think.  If you look at the cloud from the buyer side, it just makes sense to presume that they’ll eventually adopt a cloud model that offers a unified middleware/OS framework for applications, and then host components of SOA-modeled software where cost/benefit dictates.  That’s a PaaS model.  If you look from the seller side, it makes sense to offer higher layers of software because you displace more cost, easing the ROI barrier to buyers and also raising your own profit.  I think that both IBM and Microsoft have designs on making their current PaaS model the “cloud OS” of the future, which is one reason why I think both companies tend to avoid discussing that sort of thing.  Why tip your hand?  At any rate, there’s lots of potential excitement here too.