Metro Special: A question of commodity
As Ethernet prices continue to plummet, Richard Irving asks are we heading for version 2.0 of the tech bubble?
To look at a graph charting the fall in the price per unit of Ethernet capacity is to stare into the abyss, toes hanging precipitously over the edge. According to some estimates, the price of a 10G wave has fallen from around $10,000 a month to between $2,500 and $4,000 in the last 12 months alone – a drop of up to 75%. To put it another way, metro providers are now offering 10G circuits at around the same price that 1G circuits were fetching two years ago: fast forward another two years, and it is entirely feasible that 20G circuits will struggle to generate the same revenue that 1G circuits do today.
Mike Miller, chief executive of Fiberlight surmises the “price compression” (metro providers do not talk of a “crash”) thus: “It took 25 years for long distance pricing to get from the point where you were paying 75c/min for a call from New York to Washington, to the point now where that same call is free. From start to finish, the Ethernet market will reach the same measure of commoditisation in one tenth of that time.”
Moreover, he warns, commoditisation is not some dark and distant fear looming on the horizon, but a very real and present danger. “A lot of people don’t really want to talk about it, some are even in denial that it is actually happening. But the fact is that Ethernet pricing has dropped faster than any product or service that has ever come out of the telecoms stable. If we are not already there, bandwidth commoditisation is just around the corner.”
In days of old, the mere mention of the “c” word would have been enough to plunge the metro markets into a death spiral. A similar collapse in the price of long-haul capacity sparked the meltdown in telecoms assets in 2000, ushering in one of the darkest periods for carriers and network providers in generations.
Such are the phenomenal projections for capacity demand, however, that most market observers are relatively sanguine about price compression. The Boston Consulting Group, for example, estimates that more than one trillion devices will be connected to the internet by 2016 and that 45% of the world’s population will have access to broadband connectivity.
So long as sales continue to outpace the fall in prices at the current rate, metro analysts argue, the impact is likely to be muted.
Dan Caruso, president and founder of Zayo Group goes further, maintaining that it is time to lay the demons of the past to rest. “Pricing is not well understood in our industry. The view that prices are declining rapidly is an unfortunate and outdated legacy of the telecoms bust.”
If the price of metro bandwidth doesn’t continue to fall, some industry insiders accept, then households and businesses will never be able to afford the huge surge in capacity that is expected of them, putting a disastrous ceiling on growth that would be far more damaging to the industry.
“It’s similar to the dynamic at play in the iPod market”, explains Caruso. “Over time, the price of the device may not change, but you can expect to get more bits for that price. In other words, the unit price declines when measured on a dollar-per-bit-of-storage basis and that is what supports ongoing demand. The same applies to bandwidth services: Customers can achieve more capacity at a lower price per bit by upgrading to bigger pipes.”
Zayo publishes pricing trends every quarter, Caruso adds, and its experiences point to a relatively benign market where prices for a given service are relatively flat: “Prices go up as often as they go down, depending on the service you are buying”, he says.
Caruso has a point: nearly every individual in the US will use more bandwidth this year than last – and pretty much every enterprise will chew up considerably more – primarily because the cost of that bandwidth is falling so rapidly.
However, the economics of the metro market point to a slightly harsher reality where metro Ethernet can no longer command a novel premium in the market and where the march of competition continues to drive capacity rates relentlessly lower.
“This is not a market in grave peril”, reassures Miller. “But the fact is that anyone who doesn’t have a game plan to combat the commoditisation of bandwidth is going to find themselves under immense commercial pressures at some point in the future.”
Mike Sicoli, chief executive of Sidera Networks agrees: “If you don’t have a plan to drive revenue growth net of price compression, then you will be considerably behind the curve [by then end of the year],” he warns. “This is a very capital-intensive business. If you’re not growing revenue, your free cash flow is likely to be pretty stagnant and that’s a bad state to be in when you’re in a high growth market.”
Metro providers rely on revenue growth to finance new infrastructure projects. Without plenty of free cash, it becomes impossible to upgrade existing routes, let alone build new ones. And without constant investment, many metro customers, such as those that buy ultra-low latency connectivity between major financial centres, will simply disappear overnight.
Just as importantly, severe price compression could threaten to undermine the basis on which such crucial investment decisions are made in the first place. When, for example, metro providers started to upgrade routes last year, they could expect to charge a customer around $10,000 a month for 10G connectivity.
“Now, I’m only getting $2,500 a month and all of a sudden, my rate of return is nowhere near what I was originally expecting”’ complains one metro chief, who declined to be named.
A lasting legacy
Among the companies hardest hit by the pricing slump, are larger network providers with established revenues from legacy technologies: “They may be selling a lot of Ethernet, but the chances are, that they are losing a lot of SONET as well and bit-for-bit, the SONET that they are losing will be worth a lot more than the Ethernet they are winning”, explains the metro chief.
Sicoli agrees: “The companies posting the biggest net growth numbers in the market at the moment are those who started out as pure dark fibre providers and went straight to Ethernet – not those who have legacy SONET assets.” But even here, he warns, the market is maturing rapidly: “It’s going to be painful for them when some of the high value Ethernet circuits which were sold at the outset come up for renewal.”
Ultimately, he says, the point is not to try to fight customers in their drive to commoditise the metro markets, but rather, to explore existing relationships more deeply and find new ways to add value. “It should be fairly obvious that merely providing point-to-point connectivity and internet access will not generate cash flow growth forever.”
The chief executive is overseeing a major push into managed network services as well as a possible foray into cloud services in an effort to bolster revenues from enterprise customers: “A lot of companies have bought dark fibre recently and are finding it very difficult to manage. Obviously we are in a position to take on the running of their network operations at an enhanced service level for a lower cost.” At the moment, managed network services represents less than 5% of total revenue. “This is not a scaleable product that is ready to roll out across the organisation right now,” explains Sicoli, “but it clearly fosters a more entrenched relationship with customers and that is obviously attractive from our point of view.”
Fiberlight’s Miller also sees scope to offer customers a degree of value-added service: “In the end, customers won’t be looking at the price they pay for bandwidth, but rather the services that they receive off it. In this respect, the cloud is the single most important opportunity for us since the roll out of Ethernet in the first place.”
The question on every metro provider’s lips, of course, is how to monetise that opportunity. When network customers talk of the cloud, they tend to refer either to cloud computing or cloud storage. Perhaps the greatest challenge for the metro industry is to develop an Ethernet platform that establishes the whole concept of cloud transport. “This is an important issue for us this year”, explains Zayo’s Caruso. “The ultimate prize is to create the right customer experience, using existing Ethernet technology that is already out there in the network.”
Note that Caruso does not see the need to make a significant investment in new transport technology: “It’s more about establishing an end-to-end service that is easier to procure, track, pay for and control in terms of capacity needs”, he explains.
But it’s a big ask – for one thing no one is really sure what such an ecosystem should look like, least of all the clients who will ultimately finance its development. For another, businesses have yet to fully embrace the cloud, let alone articulate their transport requirements to and from it. For all the talk of virtualisation, analysts estimate that less than 5% of total IT spending is dedicated to cloud adoption.
Wider and wiser
Indeed, it is a moot point whether the network infrastructure is even capable of provisioning cloud-based services in its current form. As Gillis Cashman, a partner at M/C Venture Partners explains, there is a huge disconnect between the processing power of computers and other devices at the edge of the network and the capacity of the network to handle traffic. Cashman, whose company is a significant investor in both Zayo and Lightower, believes that this dynamic could prove to be largely unsustainable once computing shifts en masse to the cloud.
“The strain on networks will be intense. Investment in network infrastructure is way behind that on the device side of the business – and that is exactly why we are seeing such a huge drive to upgrade infrastructure now.”
A case in point is video. According to Cisco, video represented more than half of internet traffic for the first time in 2011. Yet online video still represents just 3% of all viewing. “Think how much bandwidth is being consumed even though online viewing represents just 3% of the total market – what’s that demand going to look like when the figure gets to 10%, 20% and even 50%?” asks Cashman.
Even assuming that metro providers can cope with the demands of the cloud, the threat of commoditisation will continue to rear its ugly head. “Clearly we have to figure out a way to monetise the cloud revolution beyond simple connectivity, which is susceptible to price compression”, explains Sidera’s Sicoli. The key is to establish relationships with enterprise customers, before they make the decision to move to the cloud: “That way, I get a shot at their business before they get to the data centre, where I will be pitted against four or five other providers”, Sicoli explains.
With this in mind, Sidera is looking to bolster its cloud offering with a series of new partnerships and joint ventures aimed at targeting enterprise customers. The Sidera chief is particularly keen to develop a product suite of cloud-based storage and security applications to high-end businesses: “It would make a lot of sense – but we would need to partner to get to the market,” he says.
If Sicoli sees the future further up the stack, Zayo’s Caruso is looking in the opposite direction entirely. “Our company is successful at focussing on providing fibre-based bandwidth infrastructure. Those focussed on the managed services space can be very successful but it is not our focus.”
To Caruso, the cloud will best be tapped by metro providers that develop a mechanism for delivering pure bandwidth: “Obviously we want to make it as easy as possible to consume that bandwidth, but the goal is not to deliver higher level services.” At the end of the day, he says, the infrastructure companies with the best fibre footprints covering such strategic sites such as server farms telecoms hubs and data centres will be those that ultimately enable the cloud: “It’s all about being very responsive and nimble footed and that is best done by companies who specialise in network infrastructure.”