By John Chapman, Cisco Fellow and CTO, Cisco’s Cable Access Business
As more and more cable service providers forge MVNO partnerships with mobile carriers, or build their own MNO-styled networks, the time has come to start solving mobile backhaul problems — and especially the backhaul latencies required by LTE networks — by expanding the DOCSIS specification yet again, to handle mobile backhaul on current 4G/LTE and forthcoming 5G networks.
Here’s the backstory. If it had a title, that title would be “Worlds Coming Together,” and specifically the worlds of wired and wireless network providers. Cable operators have successfully addressed the video, voice, data, home security and business services markets. Their search for the next big market has now brought them to mobile services.
Mobile carriers, faced with massive increases in consumer consumption, are working to “densify” their current macrocell topologies, which cover a radius of up to 10 miles, by moving into small cell deployments. Small cell radios will cover a substantially and deliberately smaller radius, typically of 100-200 meters, of up to 500 meters.
Meanwhile, consumers often pay their cable and mobile provider $200 or more each per month, in large part for sustained wired and wireless connectivity to the Internet. If cable operators start providing mobile services to their customers, they have the opportunity to increase their ARPU.
Clearly, though, a new network needs to be built — a small cell network. What does it take to build it? Three things, really: First, a place to mount the small cells. People call this “urban furniture,” to mean a telephone pole, lamppost, building, node, or residence. Second, those small cells will need power. Third, they’ll need a way to backhaul bandwidth — a connection, traditionally fiber. Or, DOCSIS!
The cable HFC (Hybrid Fiber Coax) plant is the only place that has all three necessities — placement, power and backhaul — because they really do have the densest wired network in the world. The HFC plant is already the RF front end for video and DOCSIS. Could the HFC plant now become the new RF front end for mobile?
(It’s worth noting that cable providers are already long-time partners to mobile carriers for macrocell backhaul services. Deepening those relationships to keep up with the pace of demand is an extension of “business as usual”).
Which brings us to DOCSIS. Cable providers are already on a path to break up their current, 500-home serving areas into 10 or so smaller serving areas. This involves taking fiber deeper into neighborhoods, with “node+0” configurations, already in play.
But taking fiber to the lamppost is just as unrealistic as it sounds, and few cable or mobile operators plan to take fiber directly to homes. It’s just too expensive, in actual and “soft” costs (like the angst that comes with digging up yards.)
Put another way: When small cells are deployed deep into the mobile network, DOCSIS will already be there. Fiber may not be.
That’s why we’ve partnered with CableLabs to develop a way for DOCSIS to grow to handle mobile backhaul — for 4G/LTE networks to start, and 5G installations as they start to emerge in meaningful ways.
From a technological perspective, it’s also notable that both cable and mobile use similar technology principles. Both are actively “densifying” their networks. Both routinely tackle the challenges associated with point-to-multipoint communications, which involves a lot of work to schedule upstream transmissions and both use OFDM related PHYs.
The mobile industry’s migration from macrocell to small cell is analogous to how cable’s HFC (Hybrid Fiber Coax) topologies are migrating from N+5 to N+0 (fiber deep) — where “5” and “0” represent the number of amplifiers in any one path along the coaxial portion of the plant.
I’ll be presenting a paper co-authored with CableLabs during SCTE Expo that describes how we can reduce DOCSIS backhaul latencies for LTE backhaul by using “pipelining” and a new message called the “bandwidth report,” or BWR. The principles proposed within it will extend to 5G.
We’ll also be demonstrating the mobile backhaul for DOCSIS concept in our booth. In essence, the intent is to create an API-based method for LTE to “pipeline” its upstream request messages to a DOCSIS CMTS in advance, so as to “work ahead” of known timing.
This matters because the minimum LTE “request-grant” time is much longer than those of DOCSIS — upwards of 15 or more milliseconds compared to 5 milliseconds with DOCSIS. What we’ll demonstrate is the low latency technique where an LTE system will make a bandwidth requests from a DOCSIS system for a specific time in the future, before the arrival of the actual traffic at the CM. This will allow the CMTS to make QoS and granting decisions earlier than it normally would.
Bottom line: When LTE and DOCSIS work together, a low latency backhaul mechanism is created for Gigabit-per-second mobile traffic. It’s a win-win that’s worth checking out!