The Sprint announcement this week of the Xohm service’s commercial launch in Baltimore (see separate item), can easily be confused with existing 3G services. I can already hear investors asking whether a WiMAX service offering 2 mbps to 4 mbps is any better than a HSDPA network which talks about a data pathway of 7.2 mbps?
The difference is of course incredible in that the WiMAX network is promising that amount of bandwidth to each customer around the clock, and that’s ALL the bandwidth that it has, and it is used for everything, voice where a network offers that (Xohm doesn’t have portable voice yet), data and entertainment services. Whereas commercial HSDPA data services are over and above existing voice channels and are a single channel of either 3.6 mbps or 7.2 mbps shared across all users talking to any given radio in one base station, and it’s for their data requirements only.
To achieve this a WiMAX network has to have around 40 mbps plus of usable bandwidth on each segment radio and usually has six segments in a Macro base station, with a total throughput of something close to 300 mbps.
Given usage patterns for broadband, that means that 1,000 to 2,000 customers can be supported directly by a large WiMAX base station and beyond that there are complexities, with many types of smaller devices used to fill out a network like Sprint’s Xohm.
Talking to streaming video company Quick Play Media this week, which has part of the Xohm streaming video contract, it gave us a perfect way to put across just how big an improvement the system represents over existing cellular services.
Quick Play is a streaming media company, and in building up our own know how on mobile TV, we have never been too kind about unicast cellular video, because it is an unnatural technology ‘ an IP protocol, RTP, laid on top of a circuit switched carrier ‘ it appears about as likely to work as IPTV did back in 2000 (alright it works now but it’s been a bumpy ride).
MobiTV has been the long standing stalwart of cellular streaming, creating its own downloadable Java application, and delivering aggregated content to it, supported by operator customers. In the past it has told us it has had to support video in 50 kbps, and that its highest bandwidth delivery in a live client was 250 kbps, and even then, this was throttled back at the base station if there was any danger of a voice call not being connected.
We discussed this will Mark Hyland, VP Marketing for Quick Play and mentioned that this particular rival, MobiTV, was the lead video provider for Sprint. ‘They ARE in 3G, but not on Xohm, because we are,’ he said, and suddenly the conversation got interesting.
At Faultline we have had to put up with people asking us for the past 4 years or so, just how do you go about providing video on a WiMAX network. Over the years the answer has changed from using an IP multicast standard like IGMP, to the wireless version designed for WiMAX called MBS, to the MXtv system available from Nextwave, and now here was a company that was about to finally do it using streaming.
‘The problem is that we are the chosen video supplier for Xohm, but now that it is about to merge with Clearwire we have to wait to see if we are going to be continuing in that role,’ he said.
But what’s the video like? ‘We have it working at Sprint in Virginia and work with another company which won the contract for supplying the adaptive bit rate. We do the feed and it is optimized for Xohm, and goes out in bandwidth of around 750 kbps. We call it HD, it looks fine on laptops, and on the Nokia 810 tablet it looks fantastic.’
And that brings us right back to our first point. This is three times faster than the fastest video offered over a cellular service which uses the same channel as the voice traffic. Even DVB-H and MediaFLO, which are multiplexed down to something way below 300 kbps, only start out life as a 400 kbps to 500 kbps encoded stream. So in one stroke the WiMAX network is getting ready to shock and awe in the video the community. However, going through the details of the announcement, video services have yet to be announced on Xohm.
This type of video will stand up pretty well against Apple iTunes video, which uses about 800 kbps for video and a bit more for quality audio. In other words cellular mobile TV will look fine on a handset a bit fuzzy on a laptop and unwatchable on a standard definition TV set. Whereas this TV will look even better on a handset, fine on a laptop, and just about watchable on an SD TV set. And given that this is less than a third of the bandwidth that a Xohm customer can expect at any given time, it makes streaming video suddenly interesting, where before it was a completely different business model, giving way to short form content due to the fact that no-one could stand to look at it for too long.
There remains one problem for a WiMAX network and that’s the fact that only 300 mbps is supporting so many customers, and that’s because services have a built in contention ratio, which can be anything from 20:1 down to 10:1. What that means is that everyone expects their 2 mbps or 4 mbps when they are online, but are not online all the time. Also they tend to do things like ask for html pages, send bulk email, even upload home made videos, but on average although they experience relatively fast broadband speed, they never use more than 300kbps across a whole period of time. So watching TV, at a constant 750 kbps would mean that if everyone did it, it would break the network. That in turn means that streaming technology can only support a VoD offering, not a TV offering which copies broadcast schedules.
Otherwise if a program came along that everyone wanted to watch, and they all switched on at once, it would lead to half of them not being able to watch it or the quality falling away (probably the former since WiMAX supports multiple layers of QoS).
So while it is still feasible that Clearwire and Xohm, once merged, might also opt for a broadcast TV option (for instance from ICO Communications which has a mobile satellite service broadcasting TV to, among other places, car seatbacks), it can still wow its rivals with a streaming service that will have real bite.
Again we still don’t have a clear picture of how the TV offering will emerge. Hyland talked about his company having 150 content providers integrated, so it can act as an aggregator, produce common video and advertising formats, but Quick Play Media has spent most of its time offering a very different kind of service. With video on a cellular network there is no QoS layer, and there are so many devices to deal with. The network speed has to be sensed, the video has to be slowed to match the device screen and resolutions and the system has to know if the processor can cope with addressing the screen, it has to ensure that it transcodes to the right video format for the player on the device. It’s a tough logistical problem.
But this is in essence what’s happening here. The Sprint WiMAX video streaming service is quite clearly not going to be given any QoS support, otherwise why would it need any adaptive bit rate services. There’s two ways to offer video over a network, make the network actively support all the bandwidth that the video requires or adapt the bandwidth of the video to what’s available. Good examples are full IPTV versus web TV. So this service it is going to be hosted off the network outside the walled garden, even if it is done with Sprint’s blessing.
Hyland explains, ‘Many of the Tier 1 US operators are frustrated at the idea of paying large up front services for content. They want to offer flat rate usage fees but that’s not consistent with paying a content owner all that money up front. For the tier 2 and 3 suppliers this is even worse and they are all looking for an outside company to put all the content rights together so they don’t have to go down this route.’
Quick Play Media also works alongside Jumptap, a company that has had investment from the same sources, which offers mobile search and advertising, and so this can also be related to the content being streamed to the customer, and adverts can be stitched together within the video stream on the fly, and this provides yet another way of paying for all those content rights and keeping some of the costs away from the wireless operators. All of this approved ‘off-deck’ activity takes a lot of pressure off operators and we may see Sprint Nextel unveiling this kind of service at the Baltimore location once it goes through its official launch next week.
‘At Cricket, a subsidiary of Leap Wireless, we are operating a similar system and we also offer a recommendation system which is personalized for each VoD viewer. There we have the viewing times going up, while Cricket is running a flat rate plan of around $5 a month. That services is only available to 200,000 handsets, but already we have 5,000 subs, and this is headed for 10% penetration, and this years the number of Cricket customers will double to around 400,000,’ says Hyland. Currently 30% of all video that is viewed on Cricket has been initiated by the Quick Play recommendation engine.
But that’s always been the trouble with streaming on cellular services. 10% penetration remains the ceiling of most services, it is always best priced at a tier 3 operator and the numbers, at a likely 40,000 if their penetration holds up, are still quite small.
Translate that into the Xohm experience, with a much higher quality of video, brand new breakthrough devices and innovative bundles and VoD services may go far higher. It will be like a network that only has iPhones on it.
Currently Quick Play Media mostly targets Motorola handsets on that service, and only recently added the Samsung Intinct and the iPhone on the other networks it’s on. In essence any device with open browsing can reach the service, which means that both iPhone’s and the Wi-fi attached iPod Touch is making up more and more of the traffic at Quick Play, which shows what hot new devices and flat rates can do for streaming uptake.