Over the past six months, there’s been a lot of talk about building a new, faster, better stronger Internet. A chunk of the discussion is driven by the proliferation of 4K video, the next new Big Thing being pushed by the broadcast industry since 3D video up and died. I think we all need to take a deep breath and get a reality check on the headaches that exist trying to tweak the status quo.
The current Internet core is fiber-based with some links ranging as high as 100 Gbps with more typical speeds in the 10 Gbps to 40 Gbps range. Moving out of the core into the edge is a mishmash of copper, coax cable and fiber connectivity, with speeds from sub megabit to gigabit or faster, depending on the technology available. Add in mobile connectivity and climbing wireless speeds in the megabit to tens of megabits, with promises of speeds up to near gigabit wireless. One of the bigger ironies here is the rapid advance of access speeds at the edge as compared to the slower move at the core. The faster we need to go at the network core, the longer it is taking to establish standards and build affordable hardware. I haven’t sat down and calculated a formal history, but it seems like it is taking longer to get 100 Gbps links into general usage than it took to move from 1 Gbps to 10 Gbps. With that track record, it will take longer still to go to 400 Gbps and/or 1 Terabit per second.
Compare core speed advances to the availability of gigabit fiber to the home, demonstrated 10 Gbps fiber speeds over existing physical plant, cable’s plans to rollout multi-gigabit capable speeds to consumers within a few years and Sprint’s talk of offering gigabit wireless speeds on LTE. The edge is speeding up faster than the core.
Now here comes 4K, also known as ultra HD. A migration to 4K in the business and consumer world is inevitable. Consumer electronics manufacturers are lining up to make affordable cameras, TVs with 4K capability are available starting at around $2000 for a 50-inch model, and broadcasters are already trialing the format with three World Cup games being broadcast in 4K. Netflix is making "House of Cards" and other shows available in 4K streaming.
Delivering 4K is simple math -- and it isn't. A 4K video stream is roughly four times as much data as today's "standard" 1080 HD video broadcast if you aren't doing compression. Broadband speeds of 15 Mbps are cited as the minimum requirement for a good 4K viewing experience, but some experts suggest service of 50 Mbps or better as a buffer since not everyone might be watching the TV on the home network and to allow for hiccups in the upstream broadband network.
Fortunately for network planners, not every household is going to buy a 4K-capable TV tomorrow and it will take time for broadcasters to migrate to the new standard, with estimates of about five years before the new format is readily available. The cable and fiber industries, as noted above, are moving to gigabit edge speeds, but new video distribution mechanisms are likely to juggle existing formats, 4K, and perhaps do some "future-proofing" by putting 8K video on the roadmap.
There are no simple answers and no quick fixes. Faster core network speed is a necessity, but the school of “big pipes fix everything” needs to go back to the drawing board. SDN and NFV provide for smarter, faster ways to configure and optimize networks beyond throwing raw speed as a fix to every problem. It will be another couple of years before we understand the impact of SDN and NFV on network operations and design, but I've heard discussions that the combination of the two can offer cost-effective solutions to distributing video. With 4K coming and core network speeds progressing at a slower pace, we'll need every new innovation we can get.