If you think about it, it's pretty clear that operator hopes for new revenue from SDN or NFV have to be realized through new services. However, most discussions on what these new services might be get vague in a hurry, and where examples are offered they tend to be fairly pedestrian -- not "new" in the sense we'd like to think.
All about the services
So what are the options for new services and the credibility of each? You can divide services into three categories: connection services that offer users the ability to move traffic between sites; connection-related services that offer data-plane enhancements (like firewalls or NAT) to connections; and non-connection services like cloud computing.
Nobody questions that non-connection services are a revenue opportunity, though whether a network operator could expect to gain a large market share from them is doubted by some. Perhaps the operators themselves wonder, though, because most of their hopes are built on one of the first two service categories.
Service providers already sell connection services, in almost every form, from IP VPNs through VLANs to wavelengths and dark fiber. I've surveyed enterprises for decades and they're pretty clear on what they want from connection service -- a real good SLA and a real low price. Decades of price competition have driven down margins on this type of service, and since enterprises typically connect branch locations, there's only limited opportunity for growth.
Elastic bandwidth, or bandwidth on demand, is the revenue-generating hope of operators in connection services today. The problem with relying on bandwidth on demand for revenue generation is that my surveys (and frankly those of many operators) show that buyers would likely use bandwidth on demand to reduce their average bandwidth, since they size their pipes to something approaching peak load now. Therefore, if users are looking to save money with elastic bandwidth, it's not a very credible revenue source.
With non-connection services, service providers theorize that adding service features like firewalls, NAT, IPsec or other CPE-like features will increase revenue. There's some credibility to this because hosted features offer the same potential benefit for economy of scale and improved cash flow that cloud computing offers. The problem is that the number of places providers can sell these features is limited -- think back to those branch office locations I mentioned above. Once current offices have hosted connection-related services like firewalls, the revenue is going to grow slowly indeed.
How about faster time-to-revenue? Another great benefit -- until they get to the revenue. If service providers shorten the time it takes to bill for a service by two weeks, they get two weeks of incremental revenue over the life of the customer. With a low rate of change for branch offices, time-to-revenue might be a good early kicker, but not a great long-term benefit.
So what's the answer? Perhaps the New IP is different from the old IP in what gets networked. Perhaps what changes IP is not how it changes VPNs and VLANs that connect sites and users, but how it is changed by a new mission, which is intra-cloud. Intra-cloud doesn't mean connecting users to cloud computing, it means connecting cloud elements to each other.
Intra-cloud services are new services, because the cloud is new. Yes, I know you think the cloud is exploding and in terms of growth that's true, but cloud computing today accounts for only about 4% of global IT spending. There is a long way to go in cloud growth, which means that network services to support the cloud have a long way to go and a lot of opportunity to generate revenue -- new revenue.
Intra-cloud services also dodge neutrality issues. If service providers want consumers to have elastic bandwidth, for example, they risk falling into paid prioritization, which in the US and elsewhere has fallen into regulatory disfavor. However, connections within cloud computing, inside content delivery networks, inside OTT services are all exempt from these rules. Service providers can offer all the dynamism they like.
The challenge with intra-cloud communication is that because the services are inside the cloud, they're invisible in a direct sense. An operator can't go out to a customer and say, "Let me sell you a dynamic intra-cloud." They have to sell a dynamic application that involves it -- applications like network functions virtualization (NFV), cloud computing (connecting application components), the Internet of Things (connecting sensors to control points and building private sensor networks to insure privacy and stability) and other emerging applications.
There are no incumbent providers for intra-cloud. There's no preferred protocol for it either, no current pricing structure, no established SLAs and no vendors have a lock on the space. This is a wide-open field and I think it could generate a boatload of revenue, though the revenue would be buried in cloud computing and NFV service sales.
The latter may be why we don't hear much about the intra-cloud opportunity because we're back to the argument that cloud computing isn't what network operators do -- they push bits. Well, if I'm right, then service providers will have to push intra-cloud bits to realize the New IP opportunity -- and they can't dodge the need to become cloud players.
— Tom Nolle, President/Founder/Principal Analyst, CIMI Corp., special to The New IP