Could Intent Modeling Save the NFV Business Case?

I’ve talked in the last two blogs about how intent modeling fits the way that SDN and NFV have to work, and also a bit about its relationship with OSS/BSS, TMF, and the ETSI models.  Today I’d like to close this series of blogs with a discussion of intent modeling and orchestration.  As it happens, it might be intent modeling that finally defines that notion fully, and also fully justifies the ISG.

When you read through the ETSI material as a (former) software developer and architect, you’re struck by one point.  The documents implicitly define software that processes data.  That point might seem trivial on the surface because, after all, isn’t that what software is supposed to do?  Actually, in a general modern sense, the answer is that it’s not.

Let’s review for a second what an intent-modeled service would look like.  It would be a series of connected/cascaded models/objects, with the top one representing the retail offering.  As you develop downward, you’d pass through some number of layers of “commercial models” that describe interconnection of purchasable entities, and finally cross a border into describing how intent was translated into resource commitments.  It’s the classic tree-turned-upside-down picture.

One could say that an “orchestrator” processes this mode, but the problem with that is already visible.  We have a separation of orchestration and management, and implicitly of deployment and the rest of the lifecycle processes.  The more logical way to approach this would be to say that orchestration is a data-driven activity, and that deployment is simply a stage in the service lifecycle.

If we presume that our service model is fully developed in the order entry process, then logically the way to deploy it would be to send the top object an “activate” command, which is an event.  That object would then run an orchestration process that would look at its own as-a-service subordinates and activate them in turn.  You’d see a cascade of these events down the tree, and a collection of responses coming back, focusing eventually on the top object.

This, IMHO, is the only way to visualize a deployment of an intent-modeled service.  All any layer of this structure knows about the rest is the set of services it consumes itself.  You commit the services at a given layer, and that commits the subordinate ones, and so forth.

Each layer in the intent model is our black box process, and so each layer presents only functionality and an SLA in an upward direction.  The model is responsible for the SLA so the model is responsible for remediation if something breaks, which includes horizontal scaling and replacement of a failed component.  Those are lifecycle actions, and so they belong in the model.

What separates this from the literal ETSI view are two points.  First, orchestration is a continuous and “fractal” process.  It is running all the time, but it’s running in response to events.  The second point is that the intent model structure describes the end-state of the service not a parameter set, and the model itself mediates processes to get the service into that goal-state.

In the intent-modeled service the “Orchestrator” and “VNF Manager” and even the VIM are all co-equal processes that are integrated into a given intent-model lifecycle state/event table.  Everything that happens is a response to conditions, which is the first test for a fully automated service process.  The blocks in the ETSI E2E are functionally valid but are not literally process elements.

One thing this could do is provide an actual mission for the NFV work to support.  From the first there’s been a strong correlation between what MANO was supposed to be doing and what something like OpenStack already does.  If you presume that the only goal of the NFV architecture is to deploy and connect virtual functions, you’ve reinvented Nova and Neutron, respectively.  But if you say that it’s the intent-modeled service, the event-connected lifecycle progression, that NFV defines, then you’ve stepped into the new world that NFV as a concept has promised.

This same approach also integrates management and operations.  A service lifecycle driven by service events has a natural event-handling relationship with both NMS and OSS/BSS.  If a service breaks, you launch a lifecycle process to remediate.  If it works, you tickle the NOC to get the underlying problem (like a bad server) fixed.  If remediation doesn’t work you have to do a billing credit by tickling an eTOM process in the OSS/BSS.  All of that is defined into the service lifecycle process by the intent model.

Presuming that intent modeling goes (forgive the pun) as intended, there is going to be an impact on the NFV marketplace.  Several in fact.

First and foremost, intent modeling cements the business case to the implementation.  If you do a PoC on an intent-modeled implementation of NFV you can tie in everything needed to show capital savings, operations efficiency gains, and agility gains.  No stranded business cases required.

Second, the relationship between SDN and NFV is firmed up, and even SDN’s role outside NFV is strengthened.  The same is true for the cloud; SDN and NFV are both mechanisms to expand the scope of cloud “orchestration” and “connection”, and so are things useful for more than they’re being considered today.

Third, there’s now a logical and valid framework in which NFV claims by vendors can be examined.  For example, if Oracle announces NFV-supporting OSS/BSS elements, how do we know that they really support NFV?  Since the relationship between operations and NFV is vague at best, almost anything can be claimed to support that relationship.  If we demand intent-modeled service with event-driven lifecycle management, we can list criteria for “support” of NFV that are meaningful.

Which leads to the fourth point, which is that this approach would separate the wheat from the chaff, NFV-wise.  We could say, for example, that if VNFs have to fit into a specific framework for intent modeling and event-driven lifecycles, then VNFs have to present specific integration points with that.  Do they?  If not, then they’re not VNFs.  Similarly, somebody with OpenStack support can’t say they are an NFV implementation unless they have the broader intent-and-event orientation.

I’d love to be able to say that this is where we’re heading.  I firmly believe that some in the ETSI process see this direction pretty much as I’ve described it here.  More probably accept the broad concept but haven’t thought through all the details.  However…everyone would not fit in that happy camp of endorsers.  There’s a lot of inertia that’s developed in NFV, and right now the momentum is not in an intent-modeled, event-driven direction.  Time will be required to turn it around.

Time we may not have.  What operators want and need from NFV has to be in field trials within nine months or so, and on its way to proving a business case, or we couldn’t hope to deploy enough NFV to change the timing of the revenue/cost crossover operators expect for 2017.  If we don’t see some strong indication that there’s at least a couple vendors who can do the right thing by around November of this year, then I think time may run out for NFV.  Which means again that we’re going to have to accept vendor-driven and even proprietary visions of NFV in order to get useful ones—like intent models.

Diving Deeper into Intent Models for NFV

I talked in my last blog about intent modeling in NFV, and today I want to look at extending intent modeling in two directions—into SDN (which is easy) and into management (which is less than easy).  I’m not going to recap the theory of intent models beyond a sentence, so if you didn’t read yesterday’s blog I suggest you do that before reading this one.

An intent model describes a “service” in terms of the characteristics it presents to its users, not by how those characteristics are created.  Thus, it’s a good statement of an abstraction for a given service or at a given interface point.  I’ve suggested that NFV could use intent models to describe both the “input” to NFV from the higher-layer OSS/BSS (a service model) and the “output” of MANO into the Infrastructure Manager(s) that represent resources.

Since NFV has to command connections be made, and since a NaaS abstraction is the obvious way to describe these connections, an intent model is almost surely what NaaS abstractions should look like.  Since SDN is a way of creating a NaaS connection set for NFV, then an intent model can describe SDN services.  In fact, better than any other approach, particularly for OpenDaylight.

OpenDaylight, with multiple southbound control options, could create a connection service using a combination of OpenFlow/SDN and legacy technologies.  To make that work optimally you’d need a description of a connection service that’s technology-neutral.  Which means an intent model.  The ODL controller could then translate intent into commitments through the southbound APIs to the real devices or management systems.  To me this is a no-brainer, to the point where you have to wonder why it’s taken us so long to get around to recognizing the value here.

The management side, as I’ve already noted, is more complex but in the end also very logical.  An intent model of a network service would include the service connection points and the “SLA”, which would describe the connection behavior to be delivered.  IMHO, it’s not much of a stretch to say that an intent model should define SLA variables, the usual stuff like bit rate, packet loss, delay, and so forth.  We could also establish a convention that every intent model defines a “service quality gauge” that would range from green through red depending on just how good/bad the service was conforming to its SLA at a given time.

The benefit of this is that a given intent model would then be defining variables that it was the responsibility of the lower-level structure (which could be a hierarchy of models or an element that commits resources) to populate.  Since an intent model would have to offer what would look like a downward policy flow to populate lower-level deployment/connection processes with SLA variables from above, there’s nothing wrong with having the flow be bidirectional.

There are implications to this approach beyond simple management variable derivations.  A “service” is now a set of linked intent models, each of which is asserting an as-a-service property set to the stuff above it.  Because every model is a service with policies, properties, and SLAs every model can be managed.

This addresses an important point in service-building, which is how you organize complex services.  Say you have a VPN with two sites, so you have three “services” in NaaS terms—two access services and one interior VPN service.  You can define this as a tree with the retail service on top and the three subordinate elements below.  The “orchestration” of this combination can be visualized easily as the instantiation of the three individual subordinate services followed by the interconnection of the three.  The former step is defined by the intent model for each subordinate, and the final step by the intent model for the top (retail) layer.

The end game here is simple.  A service is modeled and created through a series of abstractions—“objects” in modern software terms.  Each object is an intent model in its upward-facing direction and each object decomposes the fulfillment of the intent it advertises by drawing on services/intents from below.

In most cases, I think, one of these objects will have two primary service/connectivity missions.  One is to activate the services below, and the other is to “connect” them into a structure that fulfills that object’s own intent model.  Overall, this is a modeled, data-driven, approach to NFV.

And more.  If each of these models/objects has a state/event process that describes how it responds to conditions during the service lifecycle, including how it is impacted by and impacts the state of subordinate and superior models/objects, then you can describe remediation of faults and operations/management integration simply by linking processes into that model.

To me, this could overcome what could be a serious implementation flaw in NFV, which is the use of specific logic to address service lifecycle processes.  Yes, you can establish a parameter in a descriptor file to indicate whether a given VNF needs a DNS or DHCP server or a load balancer, as long as you recognize when the software is built that the parameter will be needed.  It’s hard to address new requirements without changing software.  But if you have a model-and-table-based approach with structured service objects, you can change things by simply changing the data model to invoke a different process.

I said yesterday that the introduction of intent modeling into NFV could be a major asset, a major revolution.  Like all revolutions, it could cause significant disorder.  My perception of the ISG process is that it’s proceeding in two directions at once.  One direction is augmenting the original E2E model of NFV, which I believe is not suitable as an implementation guide, and the other is trying to develop functional requirements that, if accepted, would suggest a completely different approach should have been taken.

NFV is going to have to get through this at the political level and not at the technical level.  Somebody, somewhere high up in the organization, is going to have to sit down with a small number of like-minded people and get this right, and then work out how to develop support for the change.  That support might involve some fancy footwork to paper over the fact that a lot of detail introduced so far was introduced based on the wrong approach.  It’s not necessarily going to be an easy task to change, but I think it’s essential if NFV is to be deployed optimally, and I think some operators (and even a few vendors) are seeing the light.

Intent Models in NFV: More than “Useful”

A piece of good news on the NFV front is that the ETSI ISG is moving toward consideration of intent modeling as the basis for a number of important interfaces.  I’ve been advocating intent modeling for several years now, so obviously I’m pleased with the move.  You should be too, and everyone else in both the SDN and NFV spaces, because the decision if implemented could have significant short- and long-term benefits.

In some prior blogs, I’ve suggested that both SDN and NFV must be framed around the concept of “network as a service” or NaaS.  NaaS is an abstraction of a connection service, such that it can be realized using any suitable technology or combination of technologies.  A VLAN or VPN should be a NaaS abstraction.

The reason for this is that operations evolution has to lead technology/infrastructure evolution.  You can’t have customers or customer service agents looking at a service order to decide what technology will have to be used to fulfill it.  Services have to be ordered in terms of what they deliver, and that must then be translated into how it gets delivered.  An abstract VLAN can be ordered as a VLAN and built using any combination of legacy, SDN, and NFV.

The key to making this work is the abstraction itself, the representation of NaaS.  The abstraction of a connection service has to describe the connection in terms of endpoint behaviors and service-level agreements (SLAs).  Think of it as a “description” of the service (the logical topology, which in the case of a VLAN is a “multipoint” service, and its QoS attributes, for example) and a list of endpoints.

This sort of model has been known in the industry for some time (I’ve referenced it in a blog or two earlier).  It’s called an intent model because it models not how something is done but what the “something” looks like to its consumer.  There have even been papers describing the application of intent modeling to virtual networks.  However, intent modeling was not a part of the original ETSI conception of NFV, though it was the foundation of both my CloudNFV and ExperiaSphere work.

It’s fair to ask where exactly intent modeling might fit inside the ISG’s end-to-end model, and there are a number of places.  Let’s start at the top.

A “service” is first and foremost an intent model, as my example here illustrates.  Buyers don’t expect a blueprint and a heap of parts, they want a car.  Service models should be viewed at the top as perhaps a cascade of intent models, each representing a commercial chunk of the service, presented by features and not by implementation.  This could be modeled as the TMF’s “Customer-Facing Services” or CFSs.

The challenge of intent models is that you eventually have to deliver on the intent, which means that you have to be able to decompose the “intent” into realization at the resource level.  The easiest way to visualize this is to think about a point in your service cascade where the structure below ceases to be intent-based and becomes resource-based.  In the TMF world this could be described as the boundary between Customer-Facing and “Resource-Facing Services” or RFSs.

It’s difficult to say with confidence where this boundary point is in the ETSI ISG world because they’ve put operations out of scope.  Instead, let’s shift focus for a moment and look at the ETSI structure to see where intent models might fit.

One obvious place is the top.  A request for VNF deployment should never be expressed in the form of resource-specific recipes because you don’t know where a given service would optimally locate its components (likely dependent on service locations, for example).  So the input to NFV MANO from the OSS/BSS (the Os/Ma Interface in the E2E model) should be an intent model, perhaps what we could call a “deployment order” to distinguish it from what’s above.

In fact, all of the interfaces that ETSI draws “inside” the MANO block of the E2E model should likely be visualized as intent models because in all of them you have an abstraction of behaviors on one side and realizations of the behaviors on the other.  However, the most important place for intent modeling besides the input to MANO is the VNFM-to-VIM connection (the Vi-Vnfm Interface if you’re keeping score).  This is where the service model’s realization has to be converted into specific hosting and connection processes, which means that network connectivity and NaaS are “below” the VIM (which you’ll recall I tend to generalize as Infrastructure Manager to reflect the fact that some of the infrastructure in a real service isn’t virtual at all).

In a general case, then, NFV could be considered to be a description of the boundary between CFSs and RFSs in TMF terms.  The input to NFV is a CFS representing an intent model decomposed from the original retail service (it might be that service if there are no commercial components, or it might be a commercial chunk thereof).  The VIM’s output has to be resource-specific, which says it’s a part of the “cascade” of resource models.

This raises what’s likely the critical intent-model-related question in NFV today, which is just where inside NFV the CSF-RFS boundary occurs.  If I feed CFS models into the top and get resource controls out the bottom, then we can infer that “orchestration” in general is the boundary function between the two.  But what does the border-crossing?  It’s either a MANO function or a (V)IM function.

I think it depends on how resource choices are made.  IMHO, a “service” can set policies within its intent model that would guide resource selection, meaning that a service could require things that preference one implementation over another.  It should never make selections itself because that would mean you’d have to retroject resource knowledge into service models.  If a (V)IM is a representative of, an abstraction of, resources, then unless you want there to be one logical VIM, you have to assume that RFS modeling takes place above it, with the VIM at the bottom of the tree where models turn into actual resource commitments.

I think this same approach is the way that federation of multiple providers or even SDN domains should be handled.  Each “domain” is associated with a resource model that can commit the infrastructure within.  This part of the model is (V)IM-related.  Above the (V)IM is a model that describes how services are decomposed by domain.  You could then visualize a (V)IM not as a piece of logic but rather as a set of models, which says that pretty much the whole of NFV can be visualized as an intent model and the decomposition of that model.

Which is where I think NFV should go ASAP.  Accepting some approach like this would help focus the ISG on the specific things required to make NFV work as part of a service process, and that’s the way it has to work if it’s going to deliver service management efficiency or service agility.

What Do Operators Say are the “Myths” of SDN and NFV?

Sometimes our technologies are more defined by the stories told about them than about their realities.  SDN and NFV are no exceptions, and the full scope of mythology for either would take a lot more than a single article to cover.  Fortunately we can narrow the scope of myths (and blogs) by focusing on what operators think are myths.  I’ve culled through the views of service planners in the operator world and collected what they say are the SDN/NFV tales they think are the tallest.

Number one on all the lists is the myth of the five nines.  We’d have SDN and NFV coming out our pores if we could only get them to five nines.  Some operators think this is funny/silly and others think it’s destructive, but service planners agree it’s a myth.  They make two critical points to validate their view.

First, the network business has for years demonstrated that buyers will trade off reliability for price, at least up to a point.  According to the planners, the actual target for future services is only three nines, which is where their own research say buyers will draw the line.

Second, the notion that the nine-count for SDN or NFV is somehow holding up deployment is nonsense.  The problem with both technologies is simply one of proving a business case, which in the service planners’ experience is mostly relating to proving a radical improvement in operations efficiency is possible.  They don’t think SDN hardware is any less reliable than current network devices.  They have not planned to replace major transport/connection devices with NFV functions.  Counting nines isn’t the problem, it’s counting incremental return on investment.

The second myth the service planners listed seems to come out of frustration with some aspects of both SDN and NFV standardization.  Call this one the myth of the perfect resource.  According to this myth, we have to be able to define the resource needs of a given service with great precision.  The greater the better, because if we could find that perfect match the service would be more profitable.

Wrong, say the planners.  The fact is that “perfect routes” differ only a little in cost from imperfect ones, in the scope of services being considered for SDN.  In NFV, a lot of attention to the precise resource needs of a given virtual function means that the effective size of the resource pool is much reduced, which means less economy of scale and more overprovisioning.  A highly detailed assessment of deployment needs would also be more operationally complex to administer.

Myth number three is new revenue is going to come from shortening time to revenue.  We hear all the time about service agility, but when you dig into the comments you find that the talker means “provisioning in hours rather than weeks”.  That shortening is a limited benefit because most service delays are caused by access provisioning, which can be solved only by prepositioning capacity (and that can be done without either SDN or NFV).

The longest provisioning times today are for business data services, and these services tend to get provisioned when new sites are added.  That happens relatively rarely; sites turn over at an average rate of less than 2% per year.  In any event, planners say, you can’t necessarily run glass to every building hoping a customer will pop up there, and without access you can’t change provisioning times meaningfully.

The one area of exception to this is the area of feature hosting for add-on features like firewalls, NAT, DNS/DHCP, etc. which we’ve come to associate with virtual CPE.  Planners admit that there’s likely a benefit to being able to sell users add-on connection features, but they also say that within a year or two nearly all connections have been equipped with the features they will need in the long term.  Smart planners are asking what the cost of ownership will be when the dynamic period has passed.

The fourth myth service planners cite is controversial even among the planners.  That one could be called the operations-business-as-usual myth.  According to both SDN and NFV advocates within the operators’ business, SDN and NFV can be managed in the same way as the traditional devices could have been managed.  No changes to either OSS/BSS or NMS practices will be required.  Planners divide on this one, not based on whether there’s a myth here but based on why and how it would be dealt with.

A slight majority of planners believe that it is possible that “virtual device” management practices that made SDN or NFV look like traditional equipment could work.  The problem for this group is that the qualifier is unacceptable because the consequences of being wrong would be truly dire.  This group wants to see specific service trials to prove out the model, and they don’t see results they can bank on as yet.

The remainder of planners think it would be a waste of time to try to prove the point because it’s invalid on the face.  An SDN enclave or a bunch of hosted functions are not traditional devices.  Under the skin, they cannot be managed in the same way at all.  SDN is not adaptive, it relies on setting failure modes and changing network configuration if something goes wrong.  NFV substitutes servers and virtual connections for a physical appliance, and you have to manage what you really have not what it looks like or you’ll never fix anything.

One of the most interesting things that comes out of the service planner views isn’t their myth concept but the source of the myths.  It’s become the norm to divide operators into three groups—Tiers One through Three.  While the boundaries here are soft, they do roughly correlate with the SDN and NFV socialization that’s happening.  The smaller an operator is, the more targeted the operator’s service set, the more likely it is that the operator has broadly socialized network technology changes and has hammered out an accommodation.

Service myths have the greatest impact on the Tier One operators, because those operators are the most likely to have largely independent standards and CTO processes that are budgeted and can go on for quite some time on their own, bereft of any broad support.  Tier One planners say that a lot of the mythology of SDN and NFV comes not from the vendors but from their own people, who are looking to justify (to the rest of the company) a technology they’ve been advocating.

We’ve seen some pretty interesting NFV deployments from Tier Twos, and I think that the lesson of NFV mythology is that we’ll see the light in smaller operators first, ones who have better internal cooperation.  The bigger ones should think about that one.

If SDN and NFV Change INFRASTRUCTURE What do Future SERVICES Look like?

SDN and NFV are going to change infrastructure policy, if they succeed.  I’ve blogged about that before.  They’ll also likely change the services offered by operators.  The notion of service agility as a benefit demands something to jump from (which we have; the present) and to (what we’d have to define).  I’d like to think about that future service set today.

We have three broad classes of services today.  First, we have connection services that allow us to communicate with someone (or multiple parties).  Second, we have access services that let us get to something, like the Internet.  Finally, we have private network services in the form of VPNs or VLANs.  These categories are fairly broad, and I think it’s reasonable to assume that future services will tend to work within or among the current classes of service rather than to invent new classes.

As a class of service, connection services are becoming a subset of access services, meaning that we are transforming “connection services” into “connection as a service”.  That shift is inevitable given the fact that the Internet is a connection fabric and most connection services can fit easily within it.  Traditional connection services are the least-likely to be profitable in the long term, and so one change we can expect to see is for operators themselves to under-invest in these services and accept that they’ll lose some of each (or all of some) to OTT over time.

Voice calling and SMS/IM are examples of things that just don’t make a lot of sense to spend money on as an operator.  That doesn’t mean that they’ll disappear from wireless and wireline operator inventories tomorrow, but rather that more and more of them will get subsumed into social-media and other applications.

Service-specific access is also going to slip.  I am of the view that IPTV in the U-verse sense is doomed and I think you can see that in AT&T’s quarterly report.  The whole notion of service agility demands users be prepped with enough access capacity to accept delivery of their services without physical changes being made to the demarcation.  Thus, you don’t necessarily have to give a user the fattest pipe their media can support, but you do have to give them media to support more services than they initially buy, and provide a ready means to upspeed it.

Where this trend takes is to the notion of an agile demarcation (dmark) point, a place where services can be connected for presentation to the user on demand.  The first question that SDN and NFV would have to answer is how their use would facilitate this on-demand connection/presentation.

An futuristic IP dmark could be a more sophisticated version of a gateway router (virtual or real).  A user sees an IP network as a series of accessible subnetworks, which means that it’s fairly easy to present all manner of services as simply a piece of an address space.  You’d have to make sure that you didn’t allow routing between the services for security and stability control, but that problem exists today where VPNs and Internet are provided to the same site.

One possibility here is to presume that a user has a series of IP services that are mapped to a public IP address (the popular Class A address of IPv4 RFC 1918, 10.x.x.x, is an example), which would mean that every IP connection would “see” both public Internet and a series of private services in the same connection.  Another idea would be to create tunnels with SDN that would partition the services in a way similar to that used to deliver IPTV, packet voice, and Internet over fiber.

Carrier Ethernet could be changed by this model.  Ethernet is based on Level 2 addressing, and it would be possible to give each service a Level 2 address through which it would connect.  If the service is something like VLAN that address could represent a “bridge” to an Ethernet LAN, and if the service were Level 3 (IP) the address could look like a router that’s an on-ramp to the IP subnet(s) the service uses.  You could map the same kind of service model to carrier Ethernet as to IP, which means that you could evolve both residential and business customers to the service model.

In the long run, the smart thing to do in both cases would be to give the users a fiber path that could be metered to a range of speeds, and then let the user set the total capacity of their access connection using an operations tool.  That would then set the maximum capacity of agile service connections that could be terminated on that site.

Introducing NFV into the picture changes things a bit in my view.  NFV is about hosting features not hosting connectivity, and in order to give users a feature you need to present what I’ll call a “logical service”, not a series of tunnels.  Features are either embedded in a data path, in which case they are somewhat transparent to the user except in behavior terms, or they look like something the user can address (which implies they have an address and are part of a service address space).

It’s possible to see a service set as a series of tunnels, but it’s not an easy model to manage at the user level.  Given that, I think it’s fair to presume that the future services will be presented as addresses, which means that the service dmark will be either a L2 or L3 virtual window on a world of routers and switches that are gateways to useful stuff.

Connection services in this model are simply addresses through which a connection can be created.  Private network services are the same, and access services are address windows into hosted features of any sort.  This is a kind of compositional dmark model for service evolution.  Every user gets one (or two, for redundancy) access pipes that offer an address-space (or even two—one at L2 and one at L3) window into a service spectrum.

SDN and NFV are interior technologies in this model, obviously.  The biggest change is in the service dmark device that does the composing.  It doesn’t have to be on premises (you could just terminate an access path to an interior element) and it doesn’t have to be an appliance (software virtual routers and switches would work fine).  The point is that whatever the technology, it has to be service-elastic.  Otherwise we have all this wonderful capability to turn on Service A from a portal in 20 seconds, and then watch users wait three weeks for a new access pipe to use it.

There are profound implications to all of this.  Regulators have to address the question of whether providing IP services through virtual address windows links them to the Internet and makes them a neutrality issue.  Operators have to figure out how (based in part on the regulatory issue’s resolution) how “premium” services might be offered—as Internet/OTT services or as part of a private special-service address space.  Everyone will have to figure out how to make all this stuff seem plug-and-play to the buyer.

I think it can be.  I think that the address-space windowed compositional dmark model is the logical vision for the SDN/NFV future if you look at infrastructure from the user’s perspective; if you look (virtually) down the access pipe that connects you to your carrier.  I think this is the vision that everyone has to work to, or new services are going to underperform.

Near-Term Signs and Critical Periods: SDN and NFV Before the Flex Point

If SDN and NFV are going to create market waves, the obvious question is whether vendors are going to ride them or be swept away.  Given the immature state of both technologies, there’s not a lot of clear indicators to read on that topic, but there do seem to be a few signposts emerging from the fog, and I’ll try to describe what I see on them.

First, I think it’s clear from the strong showing Huawei is making that operators (except in the US where a political decision prevents them from selling) are increasingly adopting Huawei because they’re usually the price leader.  Huawei is also driving the prices lower in deals that they don’t get.  My view is that this demonstrates that neither SDN nor NFV are seen by operators as proven strategies.  We’d be virtualizing things, not buying price-leader products, if virtualization were proven.

Second, it’s just as clear that nobody can stand up and say “SDN and NFV are never going to amount to a hill of beans!”  Operators believe in the future of these two technologies even if they can’t prove them in the present.  Juniper, for example, has done a turnaround from a couple of years ago, when they were calling “service chaining” SDN and are now calling virtual routing a VNF.  Huawei, who is winning in the current cost-cutting paradigm, is still pushing hard to be a player in SDN and NFV.  You have to be able to show relevance, at least, in an SDN and NFV future or your stuff is hit with the label of “stranded cost.”

What these two points show is that we’re in a really funny state, market-wise.  It would be incorrect to say that either SDN or NFV is a sure thing, but it’s obvious that operators are looking hard to find a reason to invest in both of them, and if the right business case emerges there will likely be a very rapid response among buyers.

What’s somewhat surprising to me is that NFV and SDN are both as much IT stories as networking stories, and IT incumbents haven’t pushed their own positions as hard as I’d have thought they would.  You can understand why Alcatel-Lucent, for example, would be charting a careful path to NFV to avoid killing more routing business in the near term than they’d gain in VNF or MANO business.  Dell, HP, IBM, and Oracle are all unencumbered by network business to lose, and all of them have server products that would augment VNF and MANO revenues in the event of a big NFV win.  Yet none of these players are really aggressive in marketing, and IBM seems almost disinterested.

I think that the lack of drama in the IT giants’ NFV positioning is responsible for the near-term focus on VNFs.  You need a truly systemic business case to drive systemic NFV or SDN deployment, and in order to get that you have to involve legacy infrastructure, operations systems, service creation, and even marketing and market targeting.  Give that list to a salesperson to sell through and they’ll respond by looking for a job with a faster payoff.  You need a strong marketing platform to push through a broad NFV success, and absent such a push from the vendors with the least to lose, we fall back on low apples.

The easiest places to push NFV and by association promote SDN are places with one of two characteristics.  First, they can be customer edge services.  Anything that is done at the service edge is customer-specific so it can be market-targeted pretty easily to manage cost/revenue exposure.  We see vCPE for that reason.  Second, they can be distributed-intelligence services which would mean mobile services and content delivery today and IoT in the future.  These services have multi-tenant value and so can be justified on at least a metro scale without pushing spending too far ahead of return on investment.

In the vCPE area, I think that the next place to watch for some serious action is the area of application delivery control and application acceleration.  These are typically presented close to the edge, and the former is closely related to the load-balancing needed for any VNFs to exercise horizontal scaling.  F5 did pretty well this last quarter, which shows that there’s a real need for the technology, and so I’d expect that we’ll see a lot of interest in this space among VNF hopefuls.

In the distributed-intelligence services space, I think we have three possible push points not yet really represented in the market:

  • IoT, which I’ve mentioned before. There really is a lot of value in IoT if you get your head out of 5G clouds and look not at how you connect the stuff but how you harness what we already can (and do) connect.  I think there are a number of vendors looking at an IoT service architecture, and I know there are operators who’d like to see it.
  • Content delivery, which is implicated in the area of greatest traffic growth and which cuts across OTT video and mobility trends to create a symbiotic architecture opportunity. We used to have a cloud CDN vendor in Verivue, and they got sold off.  True “cloud CDN” demands both SDN and NFV play a significant role, and with content as important as it is, this is a great shot.
  • Service application platform, because SDN connects service elements and VNFs deploy them, but you still need to have them for either to be useful. Logically speaking, we should have a “VNFPaaS” architecture that lets operators quickly assemble services from useful functions within a framework whose APIs and controls assure interworking and operationalization.  Right now every VNF is an island, and that’s no way to build an ecosystem.  We’d end up with the VNF equivalent of Darwin’s finches.

I think you can see the street signs, as I noted, but we’re missing that big overhead freeway sign.  For that, we’ll have to wait for Cisco to commit.  Cisco’s “let-me-show-you-just-enough” strategy for both SDN and NFV is a clear indicator that the sales champion of networking things that the current model (where Cisco is at least a major contender if not a winner) still has legs.  When Cisco starts to trot out major, and real, announcements in SDN and NFV we’ll know that the flex point, or at least the critical decision point, is near.

When might that be?  I think we’ll want to watch the signs this fall.  Operators typically do a fall strategic planning cycle that runs from mid-September into mid-November.  If they are going to do something radical in 2016 that cycle is where they’d likely show their hand.  I’ll be keeping an eye on things in that period, and you should too.

A Realistic if Unsatisfying View of the “Market” for SDN and NFV

You can hardly pick up (virtually) an online publication these days without seeing an extravagant market forecast on NFV.  I don’t have much faith in forecasts in general; they usually turn out to be aimed at validating the largest market possible because the buyers of the report are usually vendors.  NFV is particularly problematic, though, and SDN follows only slightly behind in terms of forecast difficulty.  There’s a lot of junk science out there.

Gartner made news this week by saying that neither SDN nor NFV are markets at all, which I also disagree with.  Yes, SDN is an architecture.  So is IP, and it’s a market.  Yes, NFV is a deployment option.  So is cloud computing and even Gartner agrees that’s a market.  So we can’t define our way out of facing the basic question of what the two technologies are likely to do.  We can’t fall back on avoidance.

OK, then let’s give the analysis of the SDN and NFV market our best shot.

SDN is the substitution of central control for packet forwarding using adaptive technology exchange.  If we stay with that purist definition, then SDN advances in two distinct phases.  In phase one we have the enclave model of SDN, where we apply SDN to connection problems that are limited enough to avoid the intrinsic truth that one controller couldn’t possibly software-define the Internet.  In phase two we develop or stumble on a reasonable federation model for SDN that preserves SDN value across enclave boundaries by avoiding the need to harmonize to pre-SDN network-to-network interfaces.

In the enclave SDN phase, my own model says we’re limited to penetrating about 4% of total switch/router spending.  Right now it looks like SDN would be most likely to find its place in cloud data centers first, followed by slowly advancing into metro infrastructure.  During that advance, SDN would break out of the enclaves and become a more broadly useful technology.  In phase two, my model says that it could penetrate almost 75% of the switch/router market, but this will take quite a while.

When?  The current timing my model offers is that phase one SDN cannot possibly achieve full penetration until 2018, and that by 2020 phase two would be well underway, with SDN then owning about 11% of switching/routing.  However, it’s important to note that the growth beyond that level would come largely from SDN control of virtual switch/router elements.  SDN, in short, wins by having lower OSI levels take on more of a role in grooming and fault response, and the partitioning of services and applications would then enable the use of virtual switch/routers.

NFV is a lot more complicated, primarily because it’s far from clear what specific drivers for NFV would look like.  You have to take two views—based on current benefit expectations and based on “magic bullet” expansion.  In both cases you have to look at NFV as being first a “market” in that there would be spending on NFV technology, and second as a market transformation driver that’s shifting carrier capex from network equipment to data center equipment.

Capital savings was the first NFV benefit to be cited, and most operators have abandoned it at this point.  For good reason, I think because my own model says that if that’s the justification for NFV then the maximum impact of NFV would be to shift about a percent of capex to servers.  It’s a pimple on the market posterior at that level.

If opex can be targeted, things get more interesting.  A full solution to service orchestration could justify a lot of incremental NFV spending, but interestingly enough it wouldn’t result in an enormous change in how operator capex is distributed.  In order for NFV to be operationally efficient it has to spread its wings to cover legacy infrastructure.  In doing that it creates a significant benefit even where network equipment remains the same.  What drives the transformational spending changes is the fact that the improved operations practices can then support the third benefit.

Which is new services—“service agility” in the current vernacular.  An agile and efficient operations framework would start to open new revenue opportunities for operators, which would be fulfilled increasingly through the use of servers.  In my model, this benefit-driven but unfocused drive to NFV creates a shift to servers that eventually ends up involving about a third of all carrier spending by 2025.

You noticed that I’ve not talked about what gets spent on specific NFV technology, and that was also true for SDN control.  In neither case are the expenditures a significant piece of carrier spending and my model won’t forecast something that small accurately.  A lot of this will be open source, some will be given away by vendors who will make money elsewhere.  The transformation will impact OSS/BSS spending, though, to the point where NFV-fashioned orchestration will make up almost 40% of that market, also by 2025.

My magic bullet approach does a bit better.  Recall from prior blogs that the magic bullet says that there are a small number of successful NFV applications (vCPE and mobile infrastructure, with perhaps IoT coming along).  Under this model, early NFV success is promoted in the low-apple areas where it makes sense, and these successes serve to justify near-term changes more efficiently because they limit the areas where infrastructure changes don’t pan out in immediate revenue or savings.

In this model, NFV shifts about 2% of carrier capex to servers by 2019 and that number grows to 5% by 2020.  If we could start to address IoT in respectable terms by 2016, that 2019 number would be 5% and 2020 would be 11%.  OSS/BSS impact would also be larger; 65% of that market would be orchestration-driven by 2025.

The combination of SDN and NFV would have significant impact on carrier capex.  Up to around 2020, the combination would actually increase capex slightly year over year as operators have to pre-capitalize to some degree.  By 2022 that’s turned around and there’s a decline of about the same rate, with capex stabilizing at just a bit more than 90% of current levels by 2030.

So there we are, at least as my current model shows things.  Many won’t like the lack of absolute numeric precision in my results, but I don’t think it’s possible to do an absolute forecast—not because SDN and NFV aren’t markets but because they are both alternatives to legacy equipment and their adoption will be driven by their benefits and by overall capital budgets.  I can’t forecast the latter—there are too many variables.

The most critical time for SDN and NFV is the period between now and 2018.  Any technology option that can prove its business case will drive infrastructure spending in its own direction at that point, and that which can’t prove out will be put on the back burner as operators look for better capital strategies going forward.  The moral for SDN and NFV vendors is simple.  You can’t just ride the wave here, because you’re going to have to be the wavemaker.  Otherwise something else wins.

Making Network Revolutions into Realized Revolutions

The notion that things are changing, perhaps a bit too fast for comfort, is hardly a modern phenomenon nor one confined to tech.  One of my favorite poems (Arthur Guieterman’s On the Vanity of Earthly Greatness) starts with the provocative line “The tusks that clashed in mighty brawls of mastodons…are billiard balls.”  Change, and not change for the better (from the mastodon perspective, anyway).  Our tech giants should give it a read, perhaps.

IBM was once the unchallenged leader in terms of strategic influence, with a score so high that their vision alone could drive purchase decisions.  Today they’ve fallen into a tie for third place among IT vendors and their influence and a couple bucks will get you a Starbucks.  Cisco’s score has also fallen, and so has that of Microsoft, who just generated a record loss.  In fact, all of the tech leaders are suffering, which makes you wonder just what kind of market we’re going to have in even a few years.  It also makes you wonder how we got to this point.

The root cause is likely simple.  Business technology has gotten subducted under mass market technology.  Selling expensive computers to the Fortune 500 is a decent business but nothing compared to selling smartphones to the masses of the earth.  If you look at any technology forum these days, even those supposedly focused on business IT, you find it dominated by consumer stories, consumer comments, and even consumer ads.  Absent any means of engaging the market, business IT is bound to languish.

A related issue is the death of productivity-enhancing innovations.  We’ve seen regular cycles of IT growth in the past, driven by new compute paradigms.  We ended the last cycle in about 2002 and nothing has started up since.  From that point, buyers became increasingly focused on sustaining the productivity gains they’d achieved, and the only “benefit” that a seller could offer is lower cost to achieve those goals.

You might look at this trend and see it as the dawn of the Age of IT Populism, where masses of startups rise (waving white boxes, presumably) and shake the councils of the mighty.  Well, not so fast.  While the major IT vendors are losing strategic influence and often missing revenue targets as well, they’re still sustaining market share, particularly if you take them as a group.  We might not consult IBM’s crystal ball before making decisions, but we still sign their sales orders.

It’s this contradiction that sets the tone for networking for the balance of this decade, I think.  Buyers are unhappy with their vendors, they espouse open products and open source and standards, and they fear “lock-in” and influence.  Yet they still do pretty much the same stuff as before when the PO comes out.  Perhaps the reason is an old saw; “Nobody ever lost their job by buying IBM!”  You can substitute your big vendor of choice for “IBM” here and the sense is the same.  Changing from a traditional supplier, abandoning an industry leader is a risky decision.  We’re not being paid to take risks here, says the buyer.

Which gets us back to benefits.  Suppose that private airplanes cost a thousand bucks.  Would that create an explosion in air travel?  No, but it would darn sure drive up the salaries of pilots.  What’s happening as cost of equipment and software buckle under competitive pressure is a shifting of cost focus from the cheapening capital elements of tech to skilled tech labor—professional services and integration.  We’re making something different into the limiting factor, not eliminating the limits.

In my pilot/plane example, you can see that mass-market airplanes by themselves don’t create a mass market because you can’t train people to fly them safely.  What you need to have is a product focus shift, one that takes the functional aspects of IT or network equipment for granted (computers compute, routers push bits) and concentrates on making the stuff bulletproof in terms of installation, operation, and support.

Thirty years ago, network management was said to be 18% of network spending.  In 2014 when I asked enterprises for their number, I got almost 38%.  Gear got cheaper, more gear is bought, more gear is more operationally complex…you get the picture.

Put this way, the revolutions like SDN and NFV are shooting at the wrong duck and only the cloud has a captured a glimmer of reality.  Not with IaaS, which probably consumes more tech humans than the alternative, but with SaaS.  If people want an application, sell them the application in the form they’ll use it in, not a heap of technology they can (hopefully) assemble into something that (hopefully) will do what they want.

It’s not that SDN and NFV can’t simplify things, but that we’re not seeing them as what the cloud says they’ll have to be to succeed.  This is all about network-as-a-service, not network as technology choices.  We’re back to the populist airplane thing; just making something cheaper doesn’t make it more consumable, more populist.

Almost two years ago I had a meeting with a half-dozen EU operators on the goals of NFV, and the point that was raised loud and clear was that if you wanted to do something significant with NFV, you had to address the service lifecycle.  And that’s what they say now.  One operator says “If you want ‘service agility’ then you want a complete opportunity-to-revenue lifecycle measured in days and not in years.”  Well, for all the discussions about agility, we’ve not addressed that complete lifecycle.

Part of the solution is to frame our new technology revolutions in a service management model.  The goal of all this iron is to sell services, so you need to understand how the new pieces will address that broad goal.  Making a five-hour difference in provisioning a service that’s taken you two years to frame into market terms is hardly noticeable, much less revolutionary.

NaaS is programming more than it’s anything else.  Software development principles have to guide service development if we’re going to make a massive difference in agility.  We’ve had years of good science built around agile software and now we’re trying to build agile services in activities that have nothing to do with software and have little or no participation from software people.  A software architect would never do SDN or NFV as it’s been done.

This is an easy problem to fix, too.  The software industry spent nearly two decades evolving to the modern state of modular development and as-a-service or microservice or whatever.  We have the results of all of that, and with “software” as the core of nearly every change we believe is coming in networking, it should be easy to conceptualize future services as the facile assembly of useful pre-designed components.  SDN and NFV both play a role in componentized services—connecting the components and deploying them, respectively.  What we’re missing is how the components are built, the “upperware” platforms needed to facilitate their development, and the management systems that can do drag-and-drop service creation as easily as we can do drag-and-drop development.

This is where vendors could really help.  At least three of the big NFV names (Alcatel-Lucent, HP, and Oracle) have software frameworks and skill sets.  All three have the tools needed to create an agile framework.  Let’s get to it, gang.

What Operators Think of SDN Deployment Models and What that Says about the Future

I had an interesting exchange with a planner in a mid-sized carrier, and got some insight into how network operators are seeing SDN.  Coming from an exchange with some other operators, my contact gave me a tutorial on the “models” of deployment the operators are seeing as promising.  Some are familiar, and some approaches we think of (and read/hear about) often seem to be getting discounted a bit in the real world.

The model that got mentioned first by operators in this exchange was the “vSwitch-plus” model.  In cloud or NFV data centers, SDN and OpenFlow are often used to set up the connectivity at the virtual level, through control of vSwitches.  Operators like that explicit model of setup—you connect what you say you want to connect—and so they are looking at a white-box data center where data center switches are OpenFlow-programmed as well.

One interesting point about this particular SDN application is that it’s seen primarily as a tool in reducing errors and advancing security and compliance.  Operators didn’t think the cost savings involved would excite senior management, and they didn’t expect operations costs to be materially impacted either, except as it would relate to perhaps the security/compliance area.

The second model that got some attention was the metro grooming model.  Operators made the point that it was becoming a practice to use lower-layer tunnels to separate traffic, services, applications, and users.  Physical media like fiber or Ethernet copper are examples, but they lack the granularity needed because advances in technology keep boosting the capacity physical media can support.  SDN could provide what might be called a “protocol-less tunnel”, an extension of physical media.

Where these tunnels should be supported is clear to operators too; primarily in fiber and in particular agile-optics gear.  There is already a strong interest in redefining networking as a series of parallel Level 3 universes separated at the physical level, rather than as a universally connected Level 3 world as Internet advocates might see it.  Right now this interest is focused in the metro because (logical, huh?) that’s where most of the capex is flowing.  It’s also true that metro is the focus of a lot of different services, and also a focus of net neutrality planning for operators.

Neutrality is a big reason to worry here.  If you were to build a hypothetically fully connected IP network as your metro foundation, you could (in the US and Europe in particular) find yourself defending any special service capacity against neutrality complaints because what you did would look and act much like an extension of the Internet.

There’s also an opex dimension here, but it’s more indirect.  Operators say that the opex cost per handled bit is highest at Level 3 and declines as you drop through to the OSI basement.  They also say that compartmentalized IP, particularly with topology and resiliency ceded to agile lower layers, is cheaper to run and the per-device traffic (being that of a single compartment) is low enough to make a hosted router/switch solution practical.

Some of you may recall a blog I did on this topic a while back—making the point that the network of the future could be far more virtual than we think if we compartmentalized applications and services at a virtual Level 1.  Maybe I’m self-fulfilling here, but I see some of that working its way into operator planning as this group sees it, at least.

The third model of SDN is one of end to end service separation.  The thought is that this would build out from metro commitments, meaning that operators would establish metro grooming as above, then extend it to do things like provision Carrier Ethernet using explicit OpenFlow forwarding.  These Level-1-separated services at the metro level would then be interconnected using “core grooming” to create end-to-end services.

I had hoped to see some interest in building cloud computing services this way.  It’s easy to say that you could combine the vSwitch-plus and metro grooming infrastructure choices to build a cloud data center with an application-specific network, then tunnel it to CPE in multiple sites to offer access.  This model didn’t get much attention, though.  It may be that it combines too many service areas.  Most operators run their clouds and their networks independently.  It may also be that they’re not seeing the end-game for that approach yet, and so can’t really justify diving into it.

One thing I didn’t hear from my operator source here was a goal of displacing traditional switching and routing with SDN.  I was somewhat surprised by the very focused interest, and wondered if that were simply an indication of the operators identifying low-apple opportunities from which they could build to their real goal.  No, they say—these low-apple opportunities are the real goal.  Strategic use of SDN, meaning the conceptualizing of a pure, true, SDN model for switching and routing, is simply not something they were thinking about.

I wonder whether this might not have a significant impact on SDN, but also on NFV and even carrier cloud computing.  In a way, it makes sense.  Here are the operators, saying that they are facing profit compression because of the commoditization of bit services.  Do they then re-architect bit services to offer better cost points in a technology-driven revolution, or do they just focus on the stuff that’s 1) costing a lot and 2) getting currently refreshed or capitalized?  The latter, they say.

If this is the case, then SDN needs to focus a lot more on “interworking”, both in a vertical sense with the L2/L3 stuff and horizontally from network segment to network segment.  Explicit SDN interworking beyond that accomplished using network-to-network gateway processes from L2 and L3 is essential if you’re trying to do internetworking below L2 and L3, after all.

For NFV, the challenge is that the views expressed by these operators challenges the notion of new services driving substantial NFV deployment.  This group would simply do an NFV-ish (probably more cloud-like, or agile-CPE like in the case of vCPE) low-apple implementation that’s not even particularly designed to go anywhere else.  To build NFV on a systemic scale you’d then need some overriding operations and orchestration benefit.

For the carrier cloud services, things could be really tough.  There really are no obvious low apples for cloud services.  IaaS is generalized but not much more profitable than pushing bits.  PaaS and SaaS require a specific market target that operators might have trouble even finding and later have problems addressing in an engaging way.

Maybe we’re asking too much here.  Maybe we need a technology vision of the future that operators can build toward without actually endorsing or even knowing about.  That goes against my own planning-intensive grain, but the market will ultimately decide what it’s willing to do.

The Five Stages of VNFs

VNFs, meaning virtual network functions, are important to NFV.  Without them there’s no possible business justification to be had, no matter how good our infrastructure or orchestration and management might be.  Well, we all know there are supposed to be five stages of grief.  I contend that there are five stages of VNF too, and our progress through them—as vendors and as an industry—may decide whether we can forestall the other five stages.

The first stage of VNFs is the billboard stage.  In this stage, VNF vendors eagerly seek publicity and in many cases do that by linking themselves to any vendor who can spell “ecosystem” (even if they get a few letters wrong).  The reason for this is fairly obvious; VNFs can’t deploy except as a part of a broader NFV ecosystem and it’s far from clear early on who the winners in that space might be.

Most VNF providers are in the billboard stage now, and they’re there because there’s little barrier to being a partner to the NFV masses and little they can do to get traction except as a part of either an NFV trial or a larger RFP.  Most will probably never get out of this phase, because the primo players at the ecosystem level aren’t interested in a cattle call, they want offerings they can justify integration and trial efforts on.  Many are called but few are chosen here.

The second stage of NFV is the see-what-sticks phase.  A VNF vendor enters this phase, typically from the billboard phase, when they start to understand that real engagement is going to involve a significant commitment of resources to every “ecosystem partner” they think has a shot.  Since most probably won’t qualify, this phase is really about weeding out the chaff so you can focus on the few great ones.

This is the phase where a vendor typically learns the details of operationalizing NFV, and the way those details will impact VNFs.  In many cases this will generate a significant amount of new development.  In some it may bring relief.  Some VNF vendors will realize that their “success” depends less on an ecosystem partner than they might have thought because the nature of their product (part of a business services vCPE chain, for example) allows them to deploy with little or no real NFV linkage.  Those VNF vendors will wait here a while as things develop on a broader NFV front.

Stage three is the ride a magic bullet phase.  Here, VNF vendors discover that NFV opportunities that seem real are focused on a very small set of services, justifying a small set of VNFs.  Right now, for example, the two magic bullets are mobile and business-related vCPE.  IoT could be added.  Where a VNF vendor happens to have such an offering, this offering now becomes the Road to the Promised Land, and (mixing a few metaphors) they’ll forsake all others to get their key VNF or VNFs buffed up and ready.

Some vendors, of course, won’t be that lucky.  For that group, the only hope is to find some connection between what they can do and the key opportunity drivers that are emerging.  Security, application acceleration, and even collaboration players will all want to make their stuff look like a part of the magic bullets being recognized.  Most won’t make a decisive connection.

Stage four is perhaps the most critical of all the NFV stages, the second effort stage.  It’s not that we’re asking VNF vendors to have another desperate run at the goal, but that this is the stage where everyone at every level recognizes that one magic bullet doesn’t win a war.  NFV has to be broad-based to gain enough in benefits to justify the enormous industry effort.

This is where NFV gets real, because it’s relatively easy in the current immature world of NFV specifications and implementations to make a single VNF work.  The question is whether you can do two, and this is the stage where that has to happen.  Most operators say that they’re going to expect a pathway to their second VNF about the time they start field trials on the first, and that they’ll be looking for the features that make their NFV solution inclusive during those trials.

Which brings us to stage five of VNFs, which is reckoning.  Every VNF vendor who manages to get to stage four won’t move on, in no small part because many will have hitched their wagon to the wrong star, focusing on an ecosystem that doesn’t have the breadth.  The big question for the industry is whether enough VNFs pass through this stage without washing out.  Every NFV trial, every ecosystem, won’t succeed but I suspect that if the odds aren’t way better than 50:50 even in the early stages, there’s going to be a lot of blowback.

If you’re a VNF vendor, you need to be looking at the stages one or two ahead of where you are, at the least.  The most critical stage of all, IMHO, is stage three, where VNFs will have to prove not only that they partnered with the right ecosystem vendor(s) but also that they’ve envisioned a lively market opportunity in which their VNF can play.  Next-most-critical is stage four, where a VNF provider with one story figures out that they need two good ones at the minimum.

The obvious question is where we are, stage-wise, and it’s difficult to answer that.  Every provider of VNFs has an independent vision and program, though maturity of both vary considerably.  My rough surveys suggest that two-thirds of all purported providers of VNFs have little other than a hopeful eye on the future.  One told me that “VNFs are aspirations not products”, citing the fact that there was only an immature vision of deployment.

Of the third who actually have something, the two focus areas are vCPE and mobility/IMS/EPS.  There is a pretty solid business case for agile mobile infrastructure, even out to cloud RAN or CRAN.  The interesting point is that mobile infrastructure is more a cloud application than an NFV application; the specifics of NFV are really not needed to justify early deployment and make a business case.  The problem is that it might be difficult to develop a second act from a pure-cloud vision of mobility.  Other apps require more dynamic deployment and management, and thus depend more on NFV.

vCPE is also knotty in terms of its ability to pull through full NFV.  The “edge-hosted” model offers some benefit versus standard hardware, and like mobility is could stand alone (in this case, as an agile-edge application set) without much real NFV involvement.  Again, the challenge would be in transitioning it to a broader opportunity base.  The edge features of business services are not themselves highly dynamic, and once you’ve deployed what a customer is willing to pay for, might a custom device that’s operationally simpler be a better approach?  We don’t have enough data to determine that right now.