Net Insight Offers IPTV Alternative

Net Insight AB (Stockholm: NETI-B) CEO Tomas Duffy is just one of many senior equipment company executives heading to this year's Supercomm to talk about telco video network solutions.

But Duffy's message will be going against the grain.

While the rest of the telecom industry seems to be talking about triple-play services across end-to-end IP networks, Net Insight's chief says IP isn't up to the task of effectively handling the transport of large amounts of video traffic, and that carriers with video service ambitions need to consider using next-generation SDH/Sonet technology (see Net Insight Unveils Nimbra 600).

"The IP-only story has some drawbacks. It's a fact that if you use packet technology [to transport video], you have to over-provision," says Duffy. "I'm not saying it doesn't work -- I'm saying that it causes problems. The data transport protocol in IP is not efficient. You have to over-provision, and that situation gets more acute as traffic volumes grow."

The alternative? Use SDH/Sonet to shunt the video, and other traffic, around the network, and use IP equipment at the edge. "Our approach is much more efficient and provides the quality needed for video," says Duffy. "Next-generation SDH gives the quality assurance that's needed. Then you can add Ethernet and IP devices at the edge of the network. It's all about efficiency and low cost.

"Our solution is a combination of IP with high-quality transport. What you put in at one end comes out at the other end, and it's ideal for carrying other traffic such as voice and IP as well. Our platform is IP-compatible, but carries traffic more effectively than a packet-based network."

The latest addition to the Net Insight product family, called Nimbra, is the 600 platform, which Duffy describes as "a high-capacity, next-generation SDH switch that's ideal for carriers offering video services. The 600 has an 80-Gbit/s backplane, and the Nimbra 680 is basically the same platform but with twice the slot capacity. With a full rack you can get 1 terabit per second of switching."

The product has already been displayed in Eastern Europe, but Supercomm is its North American debut. And Duffy says there are already some takers for the new product. "We have orders for the product from telcos in the U.S., and it'll be shipping in the third quarter," says the CEO, but he wouldn't say whether the orders are from existing or new customers.

If the orders are from existing customers, then that narrows down the possibilities. Net Insight's most prominent North American customer is Broadwing Corp. (Nasdaq: BWNG), which recently expanded its video network to 20 cities (see Broadwing Expands With Net Insight). Other than that, it has a customer in Canada, and then a bunch of broadcast customers, mostly in Europe, though Duffy says T-Systems Inc. is using the Nimbra platform for some media-specific applications (see Net Insight Lands Danish Deal, Net Insight Gets IPTV Contract, Ascent Media Gets Net Insight, and Net Insight Wins TV Contract).

The CEO, though, believes that carriers will eventually spot that video specialists are using Net Insight's technology, and follow suit. "The media companies have higher demands than the telecom carriers at the moment, and they are using us. That should tell the carriers something. We've been chosen because of our technology, not because we have a big brand."

A big brand such as Cisco Systems Inc. (Nasdaq: CSCO): "Cisco is pushing the all IP story, and that gives it a lot of publicity. But we're optimistic that carriers will see it makes sense to look at what we have to offer."

Cisco is indeed pushing the all-IP story, and it's not alone: Juniper Networks Inc. (Nasdaq: JNPR) also has an all-IP message for video service providers and even won a media industry contract in the U.K. recently (see Neuf Expands IPTV With Cisco, Cisco Launches Video Barrage, and BT Builds IP Broadcast Net With Juniper).

So can Net Insight break into the telco big-time and counter the IP hype machine? It has its work cut out, reckons Heavy Reading's chief analyst, Scott Clavenna. He says that, while it's fair to say that next-gen Sonet/SDH is efficient and stable and that IP networks suffer from packet loss and require over-provisioning, that doesn't mean carriers will turn to Net Insight to build their video-enabled networks.

The best technology doesn't always win, notes Clavenna, adding that Net Insight will probably continue to pick up deals from the niche media network operators for which the Swedish company's technology is optimized. "They're OK in those applications where a studio needs to move around a lot of video among production and distribution points, but there's little hope they can expand beyond that niche and into the video-for-the-home market."

Clavenna concludes: "In the residential market, IP rules because it carries all services. IP is the unifying protocol for all services, so you just have to live with it."

— Ray Le Maistre, International News Editor, Light Reading

OldPOTS 12/5/2012 | 3:12:47 AM
re: Net Insight Offers IPTV Alternative There are also effective transport protocols like ATM and POS that efficiently use/map to SONET/SDH to transport/drop to an Eth/IP edge or CPE. Also see LR 6/1 Webinar.

The BPON FTTP also offers a way to deliver the video over a separate frequency (from Eth/IP) for broadcast TV. Thus the carriers can initially avoid the risk of Eth/IP TV. Then when it proves effective (problems are solved) the Telcos can upgrade from BPON to GPON when it becomes available.

Until the franchise regulations are resolved in their favor the Telcos may not rush to IPTV. But the major Telcos do have a Plan 'B' if competition demands.


paolo.franzoi 12/5/2012 | 3:12:46 AM
re: Net Insight Offers IPTV Alternative
There is a technical error with the claims here.

The reason one would have to overprovision an IP network would be the variability of the encoding rate of any given channel encoder. Now this is a problem for any type of network depending on how good the encoder is. The model for real time video encoding could be construed as VBR. It is pretty impractical to run this VBR over a data network or a SONET/SDH network in anything but an undersubscribed manner because any quantity of bit errors will result in a customer service call.

The biggest reason to transport over IP is the relative cost per bit of Gigabit Ethernet interfaces over any equivalent SONET interface. GigE is simply priced less.

That all said, the promised fancy features are a bit aways yet at any reasonable price point. But IPTV is workable today, perhaps not with Microsoft but there are plenty of IPTV networks out there.

OldPOTS 12/5/2012 | 3:12:38 AM
re: Net Insight Offers IPTV Alternative I agree that IPTV (or packetized video) is rt-VBR, thus if it is the only traffic carried, it requires the same pipe (under-subscribed/filled) as TDM/SONET/SDH.

However, if it is Converged with other traffic using great QoS, the oversubscribed space can be filled with traffic like best effort internet. This is possible because of the significant difference in average bandwidth (BW) required for video and any of the other services. Voice uses relatively a small amount of BW, but a lot of switching/routing capacity.

"The biggest reason to transport over IP is the relative cost per bit of Gigabit Ethernet interfaces over any equivalent SONET interface. GigE is simply priced less."
But the price per delivered useable bit (information) favors the non GigE interfaces. Many proposal's with comparison experience.

Most economics analysis is counting on the mixed traffic Convergence described above.

And as ATM and POS are designed (by CSCO and others) to efficiently map mixed traffic to SONET/SDH and with great QoS to segregate the Converged traffic to avoid interference, they offer an attractive alternative.

More ATM CBR DS0s can be carried over a SONET link than TDM DS0s. It has to do with efficient SONET mapping. SONET was designed to map all prior forms of links, and TDM fills about 2/3 of the SONET bits.

Then using VBR voice over ATM, POS or GigE the number of DS0s can be increased significantly. About 3 VBRs to 1 TDM over SONET. This is primarily because only one person speaks at a time and the dead time in conversations. This dead time is removed from traffic.

But you need a great QoS, like ATM, to better separate traffic to squeeze more out of the bandwidth, the real cost of a network.

VoIP requires some more BW than VoATM (same compression). For Video this absolute BW difference is significant.


paolo.franzoi 12/5/2012 | 3:12:37 AM
re: Net Insight Offers IPTV Alternative
In theory what you say would be accurate, but in practive Video over Ethernet is cheaper.

For Broadcast Video, one is using at least an OC-12c. Using a GigE is cheaper because in theory an end office can use a number of ethernet switches to distribute the multicast traffic. The price per port of these switches is so much cheaper than the equivalent ATM interfaces its not funny. To do the same in the ATM domain, one needs to use a device that interworks ATM and IP at layer 3 or a router.

The reason is the control protocol for channel switching is IGMP. The only alternative in the ATM domain is to use SVCs. IGMP is built into almost all GigE switches for free. SVCs are hard to come by in low end devices. Worse than that the number of connections per second is about 3 per active TV at the top of every hour. SVC processing generally is not that fast as the products were not built for it.

Finally, you are looking at absolutely the wrong point of the network to determine costs. Set Top Boxes dominate cost in any large installation. Head End costs dominate in small installations. Access Equipment is the next bit. Transport is the smallest.

The biggest exception to all of this is planning for lots of VoD. We have not discussed the insanity of the numbers that arrive for that yet. Nobody is really prepared for what a high take rate of VoD is going to do to networks.

rjmcmahon 12/5/2012 | 3:12:36 AM
re: Net Insight Offers IPTV Alternative But you need a great QoS, like ATM, to better separate traffic to squeeze more out of the bandwidth, the real cost of a network.

I see this a bit differently. I see the cost of a modern optical communications network as driven by the following:

o Labor cost to install the fiber and cost to build the carrier neutral colocation facilities
o Marketing costs to convince the public why the internet is more valueable than mass media television
o Operational and management costs of the infrastructure
o Costs associated with servicing millions of customers
o Costs associated with supporting innovation which results in technology churn

I see bandwidth as the legacy item we've defined as the tradeable good. We've put bandwidth into a market system, one with monpolist mentalities holding gatekeeper roles, who in turn create scarcity out of something which technology and physics can make in abundance.

I'm struggling to find alternatives that solve the problem. I think others have struggled as well (and almost all have given up.)

How do we turn this around? I don't have answer. Does anybody else?
OldPOTS 12/5/2012 | 3:12:32 AM
re: Net Insight Offers IPTV Alternative rj, Outsource those high people costs.

Sign In