Multi-screen delivery is still new, and still highly dependent on specialized hardward infrastructure -- at least until recently. Everything required for multi-screen delivery can be done in the cloud, though doing it properly from the cloud will require some highly sophisticated software tools. CED recently got together with Envivio VP Products and Solutions Arnaud Perrier to talk about Envivio's adoption of a software-as-a-service (SaaS) model for cloud-based multi-screen delivery. [Subsequent to the publishing of this article, Envivio formally introduced an entirely cloud-based multi-screen solution called Nuage.]
CED: What’s common or typical now in cloud-based video delivery, and what would be the distinction of a software-as-a-service (SaaS) approach?
Arnaud Perrier: What you’re seeing in the market up to now is a lot of service plays from companies like Piksel, Quickplay, Ooyala, or NeuLion that require a lot of custom design and project management and development to design a custom service for an operator. Those types of approaches typically require a lot of upfront investment in terms of NRE, capex, for a headend. Sometimes there’s a lot of custom development on the apps, the clients themselves, or the DRMs.
These are typically long-term projects, months, sometimes a year before you can get to market.
Envivio comes in as a best-in-breed software provider, which allows us to come in with a complete offering that is purely based on software, as a service, either in the private cloud, or in the public cloud, where customers might want to go the pure opex approach, with no upfront investment at all.
CED: What are the differences between going a private cloud route versus taking a public cloud approach? Would you be able to evolve from one to the other? Would you want to?
AP: Both are valid approaches. We already see hybrid models.
The private cloud approach involves your engineering team, so by definition it takes a little bit longer, and it takes some upfront investment. Usually you don’t implement all the components in your private cloud.
An interesting approach might be to use Envivio Muse encoders and Halo packagers and the Guru network management system on site, with a private data center because you have access to live content sources. Then you ship that off into the public cloud, starting with the CDN, and do the apps and the billing all in the cloud – that’s a valid hybrid approach.
But we’ve seen a class of customers – probably the most interested are the Tier 2 and Tier 3 telcos. They don’t have a facility with a data center. And maybe they don’t have the engineering team at all, or don’t have the knowledge, or the capex, in which case we pushed the notion all the way. We know our software really well. We know the components, ours and from partners – CDNs or apps, for instance. We can run this for you, completely end-to-end, in the public cloud, with Amazon Web Services or Microsoft Azure or whatnot.
We’ll give you a rate sheet. You give us your content – your live sources. We’ll ingest it and deliver your services to you.
All you have to worry about is customization of the apps and the client side. We’ll send you a monthly bill, based on usage, just like big data does – just like Amazon does.
This model really brings down the barrier to entry to offer services for that class of customers who don’t have the resources or the capex to get it done.
They want to offer multi-screen services, but it’s risky. You don’t know what the return is going to be. The finance guys and marketing guys can say, let’s outsource outright, pay monthly, pay as we go.
CED: We're now several years into the multi-screen era, and yet ROI is hardly guaranteed.
AP: No, it’s not. I’m going to be controversial saying that, but the fact is that what pays the bills today is still the traditional subscription pay model, based on a gateway or a set-top box, and that’s true whether it’s cable or IPTV or satellite. Multi-screen is part of the bundle, as a constraint and a necessity to be competitive and retain customers. But in very, very few places is it a profitable service.
People see it as a cost. It’s a necessity, but it’s a cost. It’s a lot of capex to invest, and it’s growing every day with the multiplication of devices and screens.
Consumers expect they’ll be able to consume their content on any of their devices.
There are a few exceptions. BSkyB, for instance, has been very innovative with their Sky Now TV service. It’s a totally different business model. They were able to actually market a service, that is purely OTT, and purely targeted to connected devices like the Apple TV or the Xbox. It’s especially attractive for the younger demographic.
There are different packages you could do: a la carte, based on events, on a daily basis, maybe on a weekly basis, perhaps for just certain soccer games.
It’s a model that’s disruptive, but can be profitable.
Everything else is cost-based. Everyone, especially in the U.S., is trying to leverage the new advertising revenue that you might create, but I don’t think it’s there yet.
CED: What options are available for advertising in a full-cloud, SaaS multiscreen delivery system?
AP: Advertising is increasingly an important source of revenue, especially online because of the higher CPMs. When you’re talking about dynamic ad insertion on the Internet, when you compare that to traditional TV CPMs.
A customer has a choice to try a mediation feature that allows them to backfill any unused ad slots. Let’s say they have a finite inventory, and they don’t use those slots. They can backfill it with Internet ads that are targeted to users, and aggregate those users in a revenue-share model. If successful, it might actually pay for their service.
So that’s two elements disruptive in the business model. The first way is the way we charge – you get a bill for the entire service based on usage at the end of the month. Based either on CDN capacity use – the number of gigabytes, or with user services fees, based on active users. If the service becomes successful, you pay more, but hopefully the ad revenue that comes with it becomes interesting.
The big data industry has been doing this for years. We’re saying: why should it be any different for video?
At the end of the day, that’s what really makes the cloud attractive as a service – you can get started very easily. We’re offering a free trial. You can try it for 30 days, and it doesn’t cost you anything. And if works well and you like it – you get a bill at the end of the month, and you continue.
CFOs and the product guys in our customer base love this model. They see it as the lowest-barrier approach to try new stuff. They can get it on the Xbox, whatever, it’s very easy for them to get it done.
Most of the apps we’ve been involved with, you do a proof of concept for a marketing trial, you can do this really quickly with very little risk, and that’s an attractive proposition.
CED: Are there network or infrastructure prerequisites a service provider has to meet to take advantage of this?
AP: There are potentially two pitfalls. The first is access to content. There are a lot of guys who are really excited about delivering this service, but then realize, technically, it’s the cloud, they can do it, but they don’t have the rights to enough content to make it interesting.
The other pitfall potentially is with live video content there is some cost associated with ingesting all the live streams.
But at the end of the day, and you start counting pennies, you always have options to go maybe to a hybrid model. Signal acquisition costs over the corporate Internet becomes too much to bear, you go to a CDN. We’ve seen this in Latin America, where in some places they have very slow last mile speeds. That’s why CDN costs there tend to be the highest in the world. It’s not always practical there because of the costs, and that’s why you might go to a hybrid model – because of the costs. Then you buy a certain number of caching servers and some management systems, and build a CDN to get around that particular bottleneck. There are always options for working around a barrier.
With signal acquisition costs, you can choose to put your encoders at the source, if you already have your source.
CED: With the cloud, is there ever a question of service reliability?
AP: By contract, you have to deliver four nines, or five nines, and we do. We have had a deployment that had metrics using a cloud service, and the cloud service was effectively the one that never actually went down since they started, as compared to their internal infrastructure.
Why is because you can very easily in the cloud reallocate server resources. If you’re working with an Amazon or an Azure or a Rackspace, nothing is tied to any particular device. You can always move instances anyplace. Virtually it never goes down. It’s not like a traditional headend; it’s a lot more elastic, and a lot more scalable.
From a service uptime standpoint? Most of our customers would be surprised by the reliability you get over private infrastructure.
CED: Is video delivery on a SaaS basis be something we’re likely to see from a lot of vendors?
AP: We're bringing together the best in breed products. People will come to us for the same reasons people continue to use our software in their networks today – the best encoding quality, the best features for packaging, for delivering linear and non-linear content, integration with the most DRM systems, and integration with the most CDN companies and services out there.
Whether you consider those things in the private cloud or the public cloud, it doesn’t really matter, it’s the same reasons people work with us.
And then you can differentiate on the business model. We’re allowing people to be creative.
Globally, when you look at our industry, whether it’s the large Tier 1s or the smaller companies, there’s a definite trend of people wanting solutions in the cloud.
What we’re saying at Envivio is that all of our products are software based, compared to competitors such as Cisco or Ericsson. They’re talking about software defined networks, but if you did a little deeper, all they’re talking about is software management layers, or organizing devices that can be devices existing in your headend. That’s not in my opinion a very scalable and cost-effective cloud offering based on software.
CED: Them’s fighting words.
AP: Yeah, they are. We’re up for a fight.
CED: Are there any differences when it comes to handling linear versus nonlinear content?
AP: Some customers ask about how it works for premium content. They want to be sure they can uplink to a cloud provider such as Amazon. They’re concerned about bulk encryption to protect valuable content. There are solutions to do that.
On demand is not really a challenge. It’s naturally very easy to do in the cloud. It’s been done for a while.
The real innovation is the ability to do live linear content and premium live linear content, such as sports. We’re able to demonstrate that today, and that’s really disruptive in this business.
CED: Is digital ad insertion an option?
AP: We feel we’re leading the industry on DAI. We’re the only encoder-slash-packager vendor that has massive deployments with the largest MSOs in the U.S. They have DAI based on Cablelabs standards deployed on hundreds of thousands of channels. We’ve got experience with dynamic ad insertion and blackout management, which is critical for regional sports channels.
We’ve been doing it for a while; we’re very comfortable with it. We can offer capabilities like this in the cloud; it’s just a natural extension.
We’ve got a brand new product called Experience, which is a very natural product to run in the cloud, that handles individual session management, and player management directly with the clients. It handles targeted playlists for ads, time-shifted content, alternate content, cloud DVR…
Because that has to scale with the number of users and sessions, it actually runs better as a cloud service, rather than on-site – it’ll scale better.
In the online advertising world, if you look at the set-top players, whether it’s Adobe or Freewheel – which was just acquired by Comcast – or BlackArrow – a lot of the companies we work with, they all have a cloud-based offering. It’s a very natural for us to do our piece – the processing piece – in the cloud as well.