x
Video services

BigBand Lays Cable IPTV Groundwork

Digital video specialist BigBand Networks Inc. (Nasdaq: BBND) says cable MSOs can enjoy the fruits of IPTV using a simplified delivery architecture that feeds video directly to edge QAMs and completely sidesteps the cable modem termination system (CMTS).

This "bypass" approach takes the CMTS completely out of the equation, and that's a good thing when IP video is involved, according to Doug Jones, BigBand's chief cable architect.

"CMTSs have been great for data and voice for cable. What CMTSs don’t scale well for is video," Jones says.

CMTS routers, he adds, are designed to drop packets. While there are ways to recover or conceal those errors for high-speed data services and even some voice applications, "video is not at all tolerant to dropped packets," Jones explains. If video packets are dropped, "tiling," at a bare minimum, will ensue. If an entire MPEG-2 iFrame is dropped, the result could be up to a second of black screen.

"Edge QAMs don’t drop packets... We've proven that with switched digital video," Jones says.

Technical issues aside, routing video through the CMTS isn't cheap, either. Port costs for the hotly contested universal edge QAM (UEQ) sector, however, are dropping like a rock.

And the cost difference isn't trivial. Jeff Heynen, directing analyst, broadband and video, at Infonetics Research Inc. , says a CMTS downstream costs in the range of $3,500 to $4,000. That compares to $320 to $350 for one universal edge QAM (UEQ) downstream.

BigBand's answer to this is "vIP PASS," a software upgrade that enables video heading downstream to flow directly through UEQs, where packets are prepared for delivery to Docsis-certified cable modems. (See BigBand Tunes Up Cable IPTV.)

BigBand isn't alone in this area. Cisco Systems Inc. (Nasdaq: CSCO), Harmonic Inc. (Nasdaq: HLIT), and Motorola Inc. (NYSE: MOT) are among CMTS or edge QAM vendors that have pushed or proposed the concept of IPTV "CMTS bypass" architectures, as well.

BigBand, meanwhile, has added vIP PASS to its BEQ6000 universal edge QAM, as well as its Broadband Multimedia-Service Router (BMR) 1200, its recently introduced MSP2000, and the Switched Broadcast Session Server (SBSS), a product that started off as a control plane for switched digital video (SDV) but has since been adapted to handle IPTV services, as well. (See BigBand Plays It Personal .)

Under the vIP PASS process, linear or on-demand IP-based video is fed directly to the edge QAM, where packets are encapsulated with a Docsis header and sent down the operator's plant to a Docsis cable modem. The video is then passed along to the set-top, TV, PC, or to some other display device that can speak IP. In this scenario, those packets never touch the core CMTS.

Bypassing the CMTS

Cross-platform play
In addition to the added layers of personalization that unicast IPTV services can bring to cable, BigBand also thinks its approach could also enable MSO cross-platform/multi-screen strategies. Because the IP video packets are placed in a Docsis wrapper, it follows that they can be delivered to Docsis 3.0 cable modems and then shuttled to any IP-capable device, including PCs, set-tops, gaming consoles, handsets, and even some newer TV models. In BigBand's approach, the SBSS control plane alerts the edge QAM if the device making the request is a PC, TV, or set-top so that the video that comes back is delivered in the proper format.

"Being able to have low-cost entry into IPTV enables operators to explore and innovate on the new home network and all the connected devices," Jones says.

Korean MSO goes CMTS-less
So, who's giving BigBand's concept a try today? Jones says one MSO outside the U.S. is putting it in front of "friendlies" right now. He can't name the MSO, but notes that the deployment is taking place in a country where the government now requires new video service entrants to use IPTV technology.

One possibility is South Korea, where MSOs are looking at either a CMTS bypass or a "Docsis assist" network architecture, according to an industry insider who is familiar with the international cable landscape.

BigBand has been active in the South Korean cable market. Late last year it started preparing for its first SDV deployment there by integrating Alticast Corp. 's SDV set-top client software. Keumkang Cable Networks announced the deployment of BigBand's BMR in the summer of 2007. (See MSO Taps BMR.)

In the U.S., Buckeye CableSystem , which is deploying SDV with BigBand, has hinted that it might give IPTV more than a cursory glance. (See Buckeye Picks BigBand for SDV and RFoG Gets the Squeeze.)

— Jeff Baumgartner, Site Editor, Cable Digital News

psychnerd 12/5/2012 | 4:10:16 PM
re: BigBand Lays Cable IPTV Groundwork So the premise is that routers can't deliver video? He neglets to mention that all SDV implementations receive all of their video via routers. All bigband QAMs have gige ports which is where they receive their video from. Video delivered over routers. QAMs don't buffer what they receive (like routers do) so if routers are so ineffective at delivering video why does SDV work at all given that all SDV solutions are basically built on routers?

If he wants to argue cost that's fine. Suggesting routers can't do vidoe but QAMs can is vendor FUD at it's silliest given that those QAMs currently receive their video from the routers that apparently love to drop packets.

Can you please put this "informative" article in the advertising section?
SabrinaChow 12/5/2012 | 4:10:15 PM
re: BigBand Lays Cable IPTV Groundwork Jeff called it correctly; the item identified was CMTS not router. There is a certain amount of marketing to it, but who doesn't want to save millions? =o)
Jeff Baumgartner 12/5/2012 | 4:10:15 PM
re: BigBand Lays Cable IPTV Groundwork I don't think he called out all routers, but obviously had some opinions to share about the routing functions in CMTSs...but given what you say here it's certainly apparent that the economic reasons far outweigh the technical reasons expressed. But I will be asking the core CMTS guys to weigh in on this too, even if some of them happen to sell edge QAMs. Should be an interesting debate. Jeff.
goundan 12/5/2012 | 4:10:13 PM
re: BigBand Lays Cable IPTV Groundwork The whole idea behind proposing this bypass architecture is a failure of these system houses to wean themselves off the dependency on Broadcom solutions. The downstream costs are exorbitant because they use high cost low bandwidth Broadcom reference designs mindlessly, without doing their own high speed low cost implementations using off the shelf parts. For DOCSIS 3.0 DS, you can not only meet, but also handily beat DS EQAM costs if you engineer your solution right.

If you think such a solution is impossible or too hard, it is time for you to go back to EE school and relearn the basics. Its either that or they want to protect their juicy fat margins on existing DOCSIS hardware.
Cable Bigot 12/5/2012 | 4:10:13 PM
re: BigBand Lays Cable IPTV Groundwork I'd love to see the ping diagram of the communication from the view of the cablemodem and the IPTV settop, because I'm having trouble visualizing that. What the plant spectrum would look like as well has me confused.
What I'm infering from the artile is that an IPTV DOCSIS 3.0 CM starts up and signs on to the CMTS but then when a connected IPTV settop wants video it's supposed to get content from a different DS freq than what the CMTS is using. Does DOCSIS 3.0 support channel bonding that splits what is contained in the different channels? Do you say 'bond' four freqs, two go to CMTS for HSD and voice and two go to IPTV video?
If you don't do that then the CM signs on to the CMTS and maintains the US for IPTV trick mode controls to work and uses some other DS to pick up the video. This would require a 'triple play' customer to have at least two CMs in the home, one for HSD and voice (some MSOs still split this into two CMs) and another for IPTV video. Now my home cabling is really starting to look odd.
The concept sounds promissing, bypass the HE device that is overprovisioned and not set up to avoid jitter, but to me the devil is in the details at the home.

Cable Bigot
(780 Mhz DS / (6Mhz/channel) * 35Mbps/channel = 4.5Gbps = fat pipe)
SabrinaChow 12/5/2012 | 4:10:12 PM
re: BigBand Lays Cable IPTV Groundwork It looks like you have it the nail on the head with allocating certain downstreams for separate functions.

And if this is the case, home cabling shouldn't change a bit. There is no need for multiple modems and everything remains transparent to the happy end user, :).
psychnerd 12/5/2012 | 4:10:10 PM
re: BigBand Lays Cable IPTV Groundwork He did call out CMTS routers but routers are routers. To say that a CMTS is "designed to drop packets" is silly. That's a router function not a CMTS function exclusively. Routers are designed to:
1. Determine if links are congested
2. Buffer packets until the link is not congested or the buffer is exceeded.
3. If the buffer is exceed, the router will drop packets based upon orderly rules that YOU the operator determine in advance

So routers only drop packets if you design a network where traffic exceeds link capacity.

In an all QAM/bypass model you either:
1. Engineer so that video streams never exceed the QAM link interface capacity (keeping today's inefficiency)
2. Use a central server that dynamically determines the link capacity (resource availability) and requested data (session/video request)... what a surprise that Bigband sells such a server...

So the bypass premise is that:
1. It's cheaper to either over engineer a network so you never exceed the link capacity or...
2. You use an unknown central server to manage all of the IP flow requests and map them to the subset of bypass enabled QAMs that are deployed.

Oh yea, there are still lots and lots of routers installed... the only router you've removed is the CMTS. Ooops... the CMTS is still there because you need it for the voice/data service as well as supporting the next generation of return path signalling that uses DOCSIS (DSG).

Of course when cable labs defines a central bypass policy server (which they likely never will because only 2 vendors have them) I suspect the world will have moved to a unicast video model. By that time I suspect the need for multicast functionality, bonding, queuing and buffers will have evolved and the QAMs will all require chassis swaps/upgrades to add-in lame router features.

Last I checked there were 2 CMTS vendors and 30 QAM vendors. It's pretty clear which technology is harder to implement. Every other access technology has gone through this confusion that routers aren't needed at the edge (Enterprise Switches, ATM, ISDN...).

This article contributes to the confusion that feeds on the premise that 'data can't do video'. All of that messaging ignores the fact that all VOD, Simulcast and SDV uses routers to transport the data. Only at the edge is a QAM involved. To remove one router at the edge so that additional next gen QAMs can be sold 'cheaper' when all policy and resource control is sitting in proprietary vendor servers is not helping move this industry forward.

All greenfields are using unicast over IP models to deliver video. They're not limited by the number of tuners or QAM spectrum. AT&T and Verizon VOD don't drop packets as this article implies because they engineer their networks appropriately.

The sooner the MSOs move towards a unicast and tuner independant model the sooner they'll be prepared to compete with the competition on an even playing field... seen the latest Telco ads advertising 4 show recording?...

If I've not been clear, saving money on non-standard QAM features to avoid a single router at the edge is like trying to get better fuel efficiency on the Titanic. I think the industry would be better served accelerating in a new (iceberg free) direction.


psychnerd 12/5/2012 | 4:10:08 PM
re: BigBand Lays Cable IPTV Groundwork I agree with your point. I was grossly oversimplifying. There are lots of other issues like packet overhead, buffer tuning, bit error rates, in-line retransmission that come into play. Hence why I think trying to remove one router at the edge is short-sighted at best. In the end the unicast services will be based on IP end to end. At that point you need all of the tuning options (WRED not withstanding) available and bypass QAMs with their undefined central server manager will be gates to anyone launching services.

The strategy should be to separate the services from the network constraints. IP does that. Bypass in my opinion does not. The whole conversation feeds short-term confusion.

Bypass QAMs do for cable what IP enabled ATM switches did for the telcos. Can't even buy those on ebay anymore.

paolo.franzoi 12/5/2012 | 4:10:08 PM
re: BigBand Lays Cable IPTV Groundwork
nerd,

I don't disagree with your conclusion but you do make one technical error.

In general, routers do throw away traffic in networks that are not fully congested using RED or WRED.

seven
Michael Harris 12/5/2012 | 4:09:34 PM
re: BigBand Lays Cable IPTV Groundwork Weighing in on this, by quoting the article ....


"CMTSs have been great for data and voice for cable. What CMTSs donGÇÖt scale well for is video," Jones says.


Fair statement.


CMTS routers, he adds, are designed to drop packets. While there are ways to recover or conceal those errors for high-speed data services and even some voice applications, "video is not at all tolerant to dropped packets," Jones explains. If video packets are dropped, "tiling," at a bare minimum, will ensue. If an entire MPEG-2 iFrame is dropped, the result could be up to a second of black screen.


Arguably, a bit over the top. Could a CMTS/router deliver "broadcast-quality" IPTV video? Of course.


"Edge QAMs donGÇÖt drop packets... We've proven that with switched digital video," Jones says.


Yes, and water is wet too.


Technical issues aside, routing video through the CMTS isn't cheap, either. Port costs for the hotly contested universal edge QAM (UEQ) sector, however, are dropping like a rock.


This is the meaningful point in the discussion.

That said, Docsis 3.0 M-CMTS architectures help with costs for video too. With M-CMTS, the "CMTS" MAC processing is separated from the edge QAM, allowing QAM capacities to scale more efficiently. Some questionsGǪ Are there application benefits to running IPTV video through the CMTS MAC processor? Would doing so be unduly expensive in terms of hardware resources?

This really isnGÇÖt about routers or dropped packets, itGÇÖs about dropped dollars.

It is also worth remembering that, after buying into the CMTS space, BigBand exited the market. Since they no longer sell CMTS products and some key competitors do (Cisco, Motorola, Arris), seeing them put down CMTS solutions is not unexpected.
HOME
Sign In
SEARCH
CLOSE
MORE
CLOSE