By David Díaz, Systems Engineer Service Provider, Cisco
End of October, CableLabs published the latest edition of DOCSIS specification - DOCSIS 3.1. This new spec comes with the promise for cable operators to offer multi-Gbps speeds (up to 10Gbps download and 1Gbps upload) to their subscribers. And the good news is cable operators can maintain their HFC access network architectures for this purpose, in contrast to Telcos that need to deploy a Fiber to the Home solution to reach those speeds.
But since D3.1 was announced around a year ago, it has been a hot topic in meetings and discussions with cable operators. These are my first impressions after taking a look at the spec and based on conversations with customers about this topic.
Evolution towards D3.1
If we take a look at top speeds offered by cable operators today, most are around 100Mbps and only a few have announced 500Mbps services in the near future. This level of service can be covered by current DOCSIS 3.0 with no worries. Latest D3.0 24-channel and future 32-channel cable modems could cope with 1Gbps, or 10x the max speed offered today. The bottom line is D3.0 has not been taken to its limit yet. It will be the technology used in the next 2-3+ years to increase capacity, while D3.1 silicon, cable modems and CMTS gear are developed. And, in fact, it will still co-exist with D3.1 for years to come, in the same way D2.0 co-exists with D3.0 today.
Even with no time pressure, it is expected that cable operators start planning the evolution towards D3.1 in their networks from now.
Cable modems supporting D3.1 are expected to launch in an 18 months timeframe. In compliance with the specification, they will be D3.0 backwards compatible. Cable operators should evaluate their roll out in the network in order to have a future proof cable modem plant that can be easily 'switched' to D3.1 mode as needed. Of course, initial cost of these new modems when compared to D3.0 will be an important decision factor.
CMTS evolution towards D3.1 will be the next step on the horizon, but it is a critical one that is worth to start considering today. It is most likely there won't be a one-size-fits-all solution. Each operator should assess which is the optimum evolution path from their currently deployed D3.0 architectures towards D3.1 and plan accordingly. Size of hubs, availability of spectrum, need to make node splits or video convergence plans are some of the inputs to be taken into account. The availability of new architectural options such as Remote PHY, which adds to the known Integrated (I-CMTS) and Modular (M-CMTS) architectures, is also to be considered. It is no surprise that operators would like to maximize their next 2-3 years investments compatibility with D3.1.
Another key element in this evolutionary process is the HFC network. Right now, a forklift upgrade of all HFC plant to D3.1 is not foreseen. An HFC upgrade is not mandatory to take advantage of D3.1 benefits, although it will be required in several years time if D3.1 full capacity potential is to be achieved. The reason is that new D3.1 defined upstream and downstream frequency band splits need HFC active and/or passive equipment to be upgraded to support it. This kind of upgrade is costly and operators won't take the decision lightly. But it could be a good idea to start deploying D3.1 compatible - or upgradable - HFC equipment in the network, once available, to gradually replace obsolete equipment.
Summary of some D3.1 changes
Now a quick recap of some of the main changes introduced in DOCSIS 3.1:
To cope with the ever growing appetite for more capacity and to reach multi-Gbps speeds, the following technology changes have been introduced:
OFDM modulations with support of 4096-QAM subcarriers (and even up to 16384-QAM optionally), which means a 50% capacity increase compared to 256-QAM modulation used in D3.0.
New FEC mechanism called Low Parity Density Check (LPDC) that makes transmission more robust compared to previous Reed-Solomon based FEC.
Support for extended spectrum band splits in downstream and upstream to add further capacity options.
D3.1 cable modems will be greener thanks to a new low power consumption mode, DOCSIS Light Sleep, which can put the cable modem to 'sleep' most of the time when there is no traffic. This can help operators reduce the carbon footprint of its broadband service and reduce the electricity bill for the customer.
D3.1 spec includes a full chapter dedicated to Proactive Network Maintenance (PNM). The new CMTS and cable modem features and capabilities can be leveraged to help monitor cable network status. This can be an useful tool for cable operators to find and anticipate network issues that could impact subscribers' service.
Unified specification for North American and European markets (for first time).
As expected, D3.1 is backwards compatible with DOCSIS 3.0, 2.0 and 1.1. This will make migration and coexistence much easier.
There is no doubt that D3.1 is going to expand the lifetime of cable networks for many, many years. I'm eager to see what kind of services can take advantage of these massive speeds in the future. Recently I changed my cloud backup service provider and it took several days to upload my near 100GB family photo and movie clip library. With 1Gbps upload speed, it could had been done in 15 minutes… Not bad.
Hi All, Seek your advise on this. I'm exploring mpls full mesh design 4 sites (HQ, BR1, BR2 & BR3) with two PE routers. HQ & BR1 are connecting to PE1 with "vrf HQ" & "vrf BR1" and BR2, BR3 to PE2 with "vrf BR2" & "vrf BR3" likewise. ...
Hi all, I'm trying to understand what is the purpose/importance of performing a install commit after having installed couple of packages in my ASR9906. I noticed that if I perform a reload without committing, the packages will still be available afte...
CCO documentation lists out the ability to do a password recovery for eXR with a ZTP/PXE boot.
One can also perform the operation manually, like a "turboboot" for classic XR in this facinity.
this procedure will wipe out the complete system and install a...
All, From my understanding, the following route policy:route-policy TEST_EIGRP_OUT if destination in TEST_EIGRP_OUT then drop else pass endif set eigrp-metric 10000 27000 255 1 1500end-policy Should be written:rou...