Someone in the IPTV food chain had to do some kind of Total Cost of Ownership model to entice operators to its product line, but most we have talked to over the years have been suspicious of this, worrying that either they would find themselves bogged down in claim and counter claim, and have their model rubbished as too simplistic, or they would be berated by customer for giving the game away. Either way most of them never tried it until now, with browser and increasingly middleware provider Espial putting the finishing touches to its paper a few weeks prior to the IBC show in Amsterdam, in the hope of drumming up attention at the show.
Espial’s main contention is that choice of middleware, especially what Espial calls middleware, is not just about how the service looks on the screen, but it can actually save the operator huge chunks of money, quantified in the paper as a 33% saving, when calculated on a 1 million home deployment, which it says amounts to $195 million over five years.
Espial jumped out of browser only clients a while back and has extended its set top software stack and begun to add server side systems for a multicast carousel, delivering EPG, advertising, and providing content search capabilities downloaded indices, as well as customer support applications and Skin tone abilities to change the UI at the TV set to customize operator brand, changing the EPG look and feel.
What Espial is saying in this paper is that this approach works. It keeps down set top storage costs, and it also provides a future proof layer for the provision of new services and this can be measured simply by working out costs over a five year period, including swapping out set tops, launching new applications and bringing on tap new channels and video services.
What Brian Mahony, VP of marketing for Espial is doing here is comparing his new browser by imagining the types of costs that a typical operator might incur over a five year period and providing costs for an old style HTML-only browser interface with his new set top software stack.
The Espial EVO browsers is at the heart of the Siemens Myrio system, installed at Belgacom and KPN, and the NTT system, which has 280,000 copies of Espial software in fiber-to-the-home services in Japan, and which has taken the decision to roll it out to a potential market of 30 million homes.
The white paper incorporates data from one of our US partners, MRG, for forecasting average set top costs of around $200 over the 5 year deployment. This is made up of a mix of SD devices at $125 HD/DVR units averaging $375.
Espial has then added what it knows about deploying next-generation systems using its own software and concludes that operators can reduce set-top costs by up to 39% simply by having a smaller footprint, data-driven architecture which it says leads to longer life devices. The paper then says that because other middleware needs more hardware to scale and to build adapters for ecosystem/OSS/BSS components and data management, its system will save another 35% on head end costs.
Most other costs Espial has put outside the influence of middleware including the underlying broadband network, truck roll, installation, in-home wiring, access testing, product marketing and content acquisition fees. But says that its analysis addresses about 45% of the overall IPTV deployment cost for the 5 year interval including:
1) set-top box (STB) and cost of STB maintenance, upgrades, and replacement over time;
2) IPTV head-end and ecosystem, and cost to integrate and maintain over time;
3) cost to acquire and maintain IPTV middleware;
4) service innovation including costs to develop new features, applications and user interfaces;
5) network infrastructure required for packaging and delivering IPTV content across the last mile access network to subscribers.
The diagram below shows how it sees these all breaking down, with the set top by far the largest capex ingredient in the TCO of an IPTV system.
Espial then starts taking potshots at Microsoft’s approach, talking about the way the industry has rolled out System on a Chip (SoC) set tops based on Sigma Designs or STMicro chips, which although cutting down on chip costs, will do nothing to alleviate the comparatively large software footprint of monolithic IPTV middleware systems. ‘In fact, conversations with these chip vendors lead us to believe these systems will require additional memory cards beyond what the SoC platform provides, thereby magnifying the cost disadvantage of the STB.
‘Reference designs by these chip vendors also suggest that Linux based solutions could be implemented well under $100 in parts cost (for SD STBs), while solutions based on WinCE such as Microsoft TV IPTV edition, which require additional memory and computation resources, would result in higher per unit costs,’ the report says.
One of the steps in the paper is to assume that a new set top is required after 2 years, for 25% of customers, to add new features. The assumption in the Espial side of the equation is that its architecture will let its operators get away with not doing that, but instead loading a cheaper software upgrade. It costs operators up to $1 million to select, port to, integrate and test a new set top for commercial use, the paper says, and once again claims the Espial route avoids this because either there is no hardware upgrade needed or a new box will not mean any new APIs, since these are all implemented in software. Throughout the calculations Espial points to a lower cost of set tops due to its compact nature requiring less than 20 MB of memory and can be deployed for around $90 per device.
Well we won’t give Espial any more free adverts, because every other browser based system is in the process of upgrading and taking very similar steps, such as Ant Plc, and given that the Microsoft set top doesn’t use a browser (it has just introduced one now as a small part of the set top for viewing html files), not all the criticisms can be leveled at Microsoft either.
But the paper does highlight a number of the small steps that gradually the entire industry must absorb, to take the sting out of paying for set tops and their immense integration costs for each and every installation, including VoD and DRM integration and the actual details of the multicast join/channel change, which are still not standard and neither are the methods for integrating to billing and back end OSS systems and until they are, the core shape of the Espial graph will remain more or less the same.
In the end Espial claims that its approach could cut 39% from the cost of set tops, 35% from the head end, 42% saving on implementing the middleware, shave 50% from the costs of service innovation and even 5% from the network costs at it can get by with less bandwidth. We suspect that virtually every one of the 100 or so companies in the IPTV space would want to raise some challenge to the detailed view of the Espial paper, but perhaps might agree on the general tone that set tops and IPTV networks can and will get cheaper.