Altnets and neutral hosts: Are options widening for enterprise network services?

Almost all enterprises get their network services from a big telco, but we’re starting to see a kind of network populism developing, generating things like “altnets” and “neutral hosts.” Might these new players offer things that could break the big-telco hold on enterprise services, and might they create both savings and headaches for enterprise? Yes, and yes.

An “altnet” is usually an independent retail broadband provider selling services in fringe areas – an alternative to one of the big telco or MSO providers. The idea is that an altnet focuses on a small geography, an area that has concentrated opportunity (what I call “demand density”) but is too limited to justify the interest of one of the major players. These pockets usually live outside the metro areas, in deep suburbs or towns, where enterprises often have branch offices or retail stores.

A “neutral host” is a company that develops network infrastructure assets, usually assets at the bottom layer of the network stack, and then shares them across multiple retail service providers. Often neutral hosts are targeting areas where the big players would like to offer service, but where the opportunity size can’t sustain competitive overbuild. So, somebody deploys the towers and backhaul for all to share – for a price, of course. Still, total costs are likely to be far lower than competitive overbuild would generate. That means that it’s very likely that retail services would be cheaper for enterprise buyers.

Risks and rewards of altnets and neutral hosts

Cheap is good, but it isn’t everything. About three-quarters of enterprises worry about both altnets and neutral networks. The former is a concern because small players are often a business risk, and the latter because a neutral host is a silent partner that an enterprise may not be able to validate for credibility, or even see.

There are technical concerns too. “Seeing” in a management sense is the technical issue that worries enterprises the most when considering a multiplicity of access providers, and it was mentioned as a risk by every enterprise I chatted with. They all agree that having multiple altnets would complicate operations, particularly given that it’s almost certain that the altnets’ management interfaces and SLA practices would be different. If the altnet uses neutral host facilities, enterprises worry about having yet another player in the service, one who is both critical and invisible.

The one-quarter who didn’t seem concerned about these risks weren’t necessarily dismissing the issues, but they believed that the risks could be managed through service level agreements. Everyone in this group also saw some silver linings, which only 22 of the I-see-bad-things, risk-adverse, group could find. This is the stuff we really need to look at, because if neutral hosting and altnets are going to take off, they’re probably going to need reasons why customers would like the idea, not just reasons not to hate it.

The biggest potential benefit of neutral hosting that the positive thinkers saw is better broadband in thin areas. The altnet providers today are almost always companies (or even local governments) that serve a small, higher-opportunity-dense pocket in a broader rural area. Communities of as little as 5,000 households and businesses, if they’re concentrated geographically, can be profitable to an altnet. The big telcos tend not to think of cherry-picking these small-opportunity hot spots. Most such communities today are served only through wireless or satellite.

If you’re an enterprise with a fairly large geographic footprint, a “bigfoot” as one CIO described his company to me, you’d really like all your locations to have good broadband but probably can’t quite get there. Companies like remote offices to run the same applications and offer the same level of sales and support, and that’s an issue with uneven capacity and QoS in any of your smaller communities. Of the 26 who saw some good in altnets and neutral hosts, 22 cited this benefit.

Another positive thought, considered by 17 of the 26, is that altnets could spark innovation through competition. Those little towns aren’t hot targets for the local telco and darn sure aren’t going to start a competitive rush. One provider is enough overbuild. On the other hand, a few retail providers’ interest might draw two or more altnets out of the fiber woodwork, and the one with the best service for the best price would win, which means business sites in the community would also win. Even if the community decided to deploy its own fiber, it would likely do so faster, with better technology, and broader coverage at the town’s edge, if there were multiple retail providers interested. Thus, multiple altnets sharing a neutral host might be an advantage, but only if they weren’t dividing up a profit pie that was marginal to start with.

Virtual telcos on the horizon?

Risks and rewards, or at least the potential for both, seem fairly balanced here. Could we break the tie by looking beyond the tactical concerns that influence enterprise assessments? We can surely try, and a good place to look for insight is the long-term business model.

Altnets and neutral hosts both target pockets of opportunity in a broader service area whose overall demand density is low. Could neutral hosting, because it could share infrastructure across multiple altnets, create a business case in a thinner opportunity pocket? That depends on three factors: mobile trends resulting from 5G, the value of the fiber-sharing model, and competition among neutral-host providers.

Tower companies generally tell us that there’s a need for “densification,” or the improvement in cell density, in order to fulfill 5G’s expectations. That’s surely what happened with the transition to 4G, but 4G was a true experience leap forward, and 5G hasn’t shown itself to be that at all. The jury is out on this point.

Fiber sharing is a natural step in neutral hosting, but the problem there is the number of targets. There are about 800 cities in the US with a population over 50,000, and 40% of the population live in these cities. The hottest target area for neutral fiber would be the cities that are too small to be aggressively covered by the big operators, but large enough to present an opportunity. My model says that there are only about 2,500 communities in this ideal range. The biggest jump in community numbers comes when you dip to those under 10,000 in population. The question, then, is how far below the 20,000 level can neutral-host fiber be profitable to the provider? The jury is out on that too.

Could the neutral host model spread to fiber? In theory. And it’s even possible that if it did, the neutral host provider might step up the facilities they offer to altnets and even the non-incumbent operator competitors to the primary ISPs. Might they even offer a management interface, and thus be a force that could address enterprises’ big concern – operations disorder that raises the cost and risk of altnets?

That question may be the most interesting because it raises a follow-on. If neutral hosts multiply because the risks I’ve identified are addressed optimally, could they form their own “neutral federation” and become, in effect, a complete virtual telco? Could thousands of independent gas stations fill most of our tanks, rather than a few giant oil companies? That already seems the case in many areas, so why couldn’t the model work for networks? A neutral federation could create the biggest possible competitor to current telco and cable incumbents, and totally change the enterprise network services dynamic. Gas, oil, telecommunications, anyone?