To regular readers, you’ll note that I’m a little bit late with my blog for November. I’m hoping you’ll all be able to identify and commiserate with me on the reason. Have you ever identified an issue that is so important that you devote all your energy and resources to try to solve that problem quickly? Then what happens all too often? The “solution” creates more problems than it solves! I am convinced that this situation is a corollary of Murphy’s Law and trying to untangle a “bad” solution with unintended consequences can be very challenging.
Perhaps a bit counterintuitively, I think the CATV/Broadband industry may be approaching a similar situation. I just published my latest CATV/Broadband Infrastructure Amplifier Forecast and Outlook: 2015 - 2021. This forecast addresses the amplifier building blocks that are used in various parts of the HFC (Hybrid Fiber Coax) cable network. The HFC network is a point-to-multipoint architecture with content generated in a central location and then distributed to many endpoints. As the name implies, the network is a mix of fiber and coaxial cable transport architectures. The point in the network where the signal transitions from optical transmission on fiber to RF transmission is the optical node. In Fiber to the Home (FTTH) applications, the network “plant” is fiber. This type of CATV/Broadband network offers definite advantages from a speed and capacity standpoint, but in many cases the cost to pass a home with fiber is prohibitive and operators scaled back their deployment plans several years ago.
Here’s where things get interesting. The CATV/Broadband industry has used Nielsen’s Law as a rule of thumb to help with capacity planning. The “law” was an anecdotal observation of internet connection speed versus time. From 1983 to the present, the curve has shown a remarkably consistent 50% annual growth rate. It’s been so consistent that the industry uses a variation of that analysis to forecast traffic and network capacity. This orderly process was disrupted in 2012/2013 when Google introduced their Google Fiber initiative to bring Gbps speeds to their targeted areas. This upset the apple cart for the cable industry, because this speed/time combination was significantly earlier than their planning curves predicted.
Fearing that they would lose the “maximum advertised data rate” battle to fiber networks, the cable industry responded by quickly developing and ratifying the DOCSIS 3.1 specification. This specification is a big departure from earlier versions with its use of OFDM modulation and multiplexing schemes, new forward error corrections and allows for higher frequency bands with increased bandwidth. The goal was to achieve 10 Gbps downstream (to the end user) and 1 Gbps upstream (from the end user) data rates to be compatible with fiber, at least in the downstream (fiber can have symmetrical data rates). The table below shows some of the variations of DOCSIS 3.1 and the performance versus the previous specification (DOCSIS 3.0). The resulting performance is impressive.
Compound semiconductor device manufacturers that participate in this market were very excited about the prospect of an update to the plant infrastructure, but here is where we may be seeing the solution not being what we expected. One of the “must haves” for the new specification was backwards compatibility with the earlier infrastructure. Many operators had recently invested significant amounts into infrastructure upgrades to the DOCSIS 3.0 specification and were leery about getting return on their investment. This eliminated the need for massive infrastructure upgrades, as operators are able increase performance is stages.
As network traffic continues to climb quickly, enabled by DOCSIS 3.0 and 3.1 deployments, this is starting to strain the transport network that has to aggregate all this traffic. As a result, fiber is pushing deeper into the HFC network. More fiber means less coax and less need for RF components. This “success” of DOCSIS 3.1 has spawned “Full Duplex DOCSIS” as a means of making the downstream and upstream data capabilities symmetric, like fiber. This idea is in its infancy, but there is a lot of development effort aimed at solving the challenges. This concept requires a passive fiber network with no directional/active devices. The interesting question will be whether operators can monetize the higher speed upstream path. If they can, this may change the fiber business model and make FTTH deployments more attractive.
If that is the case, the net result will be the cable industry developed the DOCSIS 3.1 specification that was so successful that it eliminated the need for RF components! That will be a great solution resulting in very bad results for the RF CATV industry. I’ll be monitoring this development closely!
If you are interested in learning more about DOCSIS, the developments and trends and the RF amplifier forecasts for this industry, please feel free to contact me about purchasing these reports or becoming a subscriber of the Advanced Semiconductor Applications Service.