Commoditization: Sometimes Less Really Is Less
As cost per bit keeps dropping, and commoditization continues, we might find ourselves in a precarious position on bandwidth.
Networking is pervasive today because cost per bit has plummeted. According to the FCC, a household with 16 times the bandwidth of a "fast" data trunk just 20 years ago doesn't even qualify as being a "broadband" customer today. Times have changed, and what has changed them is commoditization.
It's a good thing, right? As a buyer, I would say it certainly is -- but price commoditization that's not compensated by lower network costs can threaten broadband services, and maybe sooner than we think (see related post, How 2 M&A Drivers Shape Networking Overall).
Price commoditization started with packet switching. Old data services gave customers fixed, constant, bandwidth. A T1 line was 1.544 Mbps, all the time. Back in the '60s Rand Corp. did a study that proved that data traffic was inherently bursty, and so you could fill in the valleys of one stream with the peaks from another to get two or more flows to share the same capacity. IP is a packet protocol, and the Internet a packet network, and most of our cost-per-bit revolution was based on the packet.
Then we had fiber. There's a limit to how much data you can punch across copper wire, particularly if you want to connect across any distance. Fiber optics promises tens, hundreds, and thousands of miles of transport, and with wavelength multiplexing we can put multiple transport paths on the same strands of glass. Fiber transport reduced the cost of transmission significantly, and it provided the second wave of cost advantages that kept operator cost per bit falling at pace with retail price declines.
Scrounging Around for Savings
Packets and fiber built our broadband services. What we call "over the top" (OTT) services like streaming video are all products of commoditization. If consumer cost per bit were higher, then these services wouldn't be as attractive. OTTs have taken advantage of commoditization, leaving most operators feeling like they've been taken advantage of along the way. The only remedy for operators is further cost reduction, and the big question has been where to find it.
Operators are using network functions virtualization (NFV) and software-defined networking (SDN) to attack vendors on their high switching/routing equipment prices, but most will acknowledge that this approach doesn't offer much hope of enduring savings. Competition among vendors, particularly because of Huawei, has given operators leverage in negotiating those prices, and operators think that capital equipment savings won't have enough impact to sustain profit margins. For more savings they have to look at operational expenses.
Opex is a vast, complicated, mess. Network management, service management, business management, and operations management are all woven into a web of tools and "craft costs," or human expenses. Despite, or maybe because of its complexity, operators say opex has gone from a third of total cost of ownership (TCO) 20 years ago to over half of TCO today, and they expect it to climb to two-thirds or even three-quarters of TCO by 2020. Opex savings has turned into a prime objective for NFV because the opportunity to save is large, and getting larger.
The challenge is actually reaping those benefits. Neither SDN nor NFV offer any management standards improvements; the whole area is out of scope to the primary work of the standards people. Insiders think that NFV and SDN might even increase network complexity, and that would almost certainly increase opex. The fact that the standards for SDN and NFV don't address management has given vendors air cover to dodge the complex management issues, and early SDN and NFV trials have done little to even validate what opex might be.
Staving Off Starvation
The operators themselves are divided as far as where operations efficiencies might be improved, and more so on the "how?" Some want to trash the traditional operations support system/business support system (OSS/BSS), while others want to build it better and envelope SDN and NFV technologies. Whatever approaches operators take, no single technology trend is guaranteed to support their paths. SDN and NFV benefits may come more from integration services than from products. Ericsson, the network vendor with the most focus on operations and related professional integration services, reported that this quarter OSS/BSS and related integration services were a bright spot in growth and profit margins.
Meanwhile, some of the commoditization tension has entered the public policy arena. Net neutrality is really about preserving the current Internet business model, which OTTs need and which operators hate. Globally, OTTs are winning because consumers see the Internet and OTT services as critical parts of their lives. However, profit pressure on operators, if left unabated, could reduce investment in Internet and mobile access and eventually starve OTT applications (and their users) for bandwidth.
"This can't go on" is a popular refrain among network operators, and most say their revenue/cost-per-bit curves will cross over in just a couple years. The best indicator on whether commoditization will continue, or whether it will starve us all for bandwidth, may be how the FCC decides peering complaints. If operators are able to charge big OTTs for connection, as a number have done for Netflix, that may give operators a bit more headroom. But buyers and sellers are always in conflict over price and profit, and commoditization has been a key element in the success of the Internet. It's not going to go away quietly.
Follow Tom Nolle on Google+!
Tom Nolle on Google+