News
TORONTO – MSOs are beginning to offer a 100 Mbps tier, frequently for $150 per month. Given the costs and the take rates, operators should entertain the notion of cutting that monthly fee almost in half.
The notion popped out at a presentation at the SCTE Canada Summit, given by Jason Lowe, director of access network engineering at Clearcable Networks, on the experience of smaller operators installing DOCSIS 3.0 (D3) technology.
Lowe’s key points were to plan well in advance for what happens after you install D3, because there are some unexpected results.
For example, headend wiring is typically based on a device output level of about 60 dBmV. But with D3 and an edge QAM (eQAM) with four channels per port, output levels will be closer to 52 dBmV. Eventually, an operator is likely to split that to 8, and then 16 channels per port; at 16 channels, the power level drops to 45 dBmV.
Which suggests that any operator planning to deploy D3 should consider rewiring. Figure out how channels will be moved long before they are moved for channel alignment, Lowe recommended.
In a separate presentation, Steven Krapp, director of C4 CMTS product management at Arris, noted the same phenomenon and recommended that the number of splitters and combiners between the CMTS and the transmitters be reduced.
Channel realignment is likely to lead to moving the DOCSIS channels. Not that big of a deal, Lowe said, until you realize that if that’s done, then the operator must also replace all of the traps in its system. (Lowe wryly noted that no matter how far in advance trap replacement was planned, something always came up so that the process invariably was scheduled in the middle of Canadian winter).
Deployment costs are high, Lowe said, and he showed the following table:
|
High-Level Costs |
|
Notes |
|
Upgrade to mCMTS, including CMTS hardware, eQAM and headend installation |
<$150,000 |
Total headend capital expense |
|
Budgetary DOCSIS 3.0 modem cost |
~$100 |
|
|
Number of DOCSIS 3.0 subscribers per CMTS |
<$2,000 |
Total number of supportable subs based on 24 secondary channels and 25 Mbps and 50 Mbps packages |
|
Total capital expense per subscriber |
~$175 |
|
As an aside, Lowe suggested that given the very low adoption rate of 100 Mbps service, and given the relatively low capital cost per subscriber, operators could consider offering the very fastest tier at $80 a month or so.
A fringe benefit, Lowe observed, is that if you make the service on the lightly used secondary D3 channels more attractive, the migration of subscribers from primary D2 channels will alleviate traffic congestion there.
“No one seems to be thinking of this,” Lowe said. “Even if you throw $300 at each sub, the return on investment is still pretty good.”
Krapp recommended service groups of 400 to 500 for downstream, but only 200 to 300 for each upstream channel. When recombining service groups, he said, it’s important that attention is paid to upstream signal-to-noise ratio (SNR).
D3 equipment is denser than DOCSIS 2.0 equipment, Krapp noted. That may affect heating/cooling in the headend.
Another unexpected challenge is the TCP stacks in subscribers’ PCs. Traffic patterns suggest that PCs with Windows XP make up just under 60 percent of the installed base of computers.
Windows Vista, Windows 7 and Mac OS all have TCP stacks set for a maximum download speed in excess of 100 Mbps. Not Windows XP. The TCP stack in Windows XP caps it at 31 Mbps.
It is an obvious problem if you want to offer a 50 Mbps tier, but 60 percent of your potential customers will not be able to take advantage of it.
The solution is actually a simple change in the Windows XP registry, but the downside, Lowe observed, was having to walk customers through the task, and heaven help any operator if anything goes wrong with their customers’ computers any time after the process is performed.
The alternative is crafting tiers/packages that take into account the potential customer’s PC/operating system.


