How to support person to person video calling in the network
Social and behavioural changes can drive technical innovation. With SMS we saw the phone move from the ear to the palm of the hand. At the device level, UI and text input became vital. In the networks, a messaging infrastructure sector was born, and continues to mutate today along with the move to IP and app-based messaging usage.
As usage moved from audio to visual, and graphics became more important, so screens moved to colour and changed shape and size. UIs changed out of all recognition. This shift created clear winners and losers in the device sector, just as it had in the network space.
The popularity of video streaming on mobile, actually a surprise when it really took off despite a long gestation, brought about networks that required media gateways, optimisation, re-encoding on the fly, ad insertion, QoS controls and policy interfaces – a whole range of network elements required to support and attempt to monetise the fact that, give the content, given the device, given the bandwidth, users will watch video on their phones, and lots of it.
So, will person to person video calling take off? We just don’t know, do we? We can see the barriers. There’s little interoperation between OS and devices, never mind networks. There’s the social element – a time and a place for video calling perhaps? But it’s too tough to call. The first handset to get something right (the T9 of mobile video calling if you like) will drive uptake and usage. As will, it could be argued, some replication of the interoperable nature of voice calling and SMS. But if it does take off, what then for the mobile operator?
The above are just a few thoughts, spewed out as context for the following guest post from Manish Singh at Radisys, as I think it’s good to attempt to make the link between technology and social and behavioural change. The piece is a good deal more practical than my ramblings and outlines the elements that operators will need in the network to support what may yet be a social transformation. (Keith Dyer, Editor, Mobile Europe)
Will iOS6 finally kickstart the era of mobile video calling?
We’ve been talking about conversational video on mobile for years. We’ve had the ‘year of mobile video’ as many times as we’ve had the ‘year of mobile advertising’. The launch of iOS6 this month could be about to change things as it allows for the first time FaceTime to run over 3G and 4G cellular networks.
Apple said that iOS6 has been downloaded onto 100million devices. Not all of those will be cellular-enabled (particularly the iPads), but that still represents a significant number of users that will have access to peer-to-peer video calling over mobile networks.
With so many devices in the market, the impact of FaceTime on operators’ 3G and 4G networks will be significant. And Apple’s rivals will no doubt look to ramp up marketing on similar service offerings to take advantage of the trend. We could well have mobile users viewing video calling as a primary capability of their smartphone or tablet.
And that may be in lieu of operator-owned voice minutes.
This is serious. This is a dramatically increased volume of data traffic on the operators’ networks, without the commensurate increase in revenue. Those on the infrastructure side of the mobile industry have been talking about this threat for a while now, but there’s no doubt that conversational video is the thing that makes it a reality.
How can operators manage this traffic cost-effectively? How can operators deliver the required user experience to keep their subscribers happy? And how can operators ensure that their networks can scale as the popularity of video calling takes off, beyond the Apple phenomenon?
Core components of video calling
Core network elements will bear the brunt of this increased volume of traffic. They will be responsible for two crucial functions:
- Minimizing the operational expenditure inherent to transporting large volumes of premium data
- Providing operators with the platform, capability, flexibility and awareness to begin to monetize mobile video
Fundamental to this will be the following elements:
Serving Gateway: Acting as the anchor point for the core network makes the SGW the workhorse of the LTE network. Being the anchor point means the SGW must be capable of managing mobility as a FaceTime user moves through the network. This is a very different scenario to a user downloading emails and a significant amount of capacity is required.
Packet Gateway: The packet gateway provides connectivity between the device and the external network where the video server resides, but that is not its only job. The PGW can perform a number of additional tasks including policy enforcement, packet filtering, billing and potentially lawful intercept. As a result, these platforms must be capable of doing some heavy lifting with a lot of packets.
DPI Platforms: LTE, due to its increase in available bandwidth, will carry a wider range and volume of services (data, voice, video) than any previous architecture. In order to be more effective, LTE will require DPI-based network probes capable of providing real time data for the PCEF. The DPI platforms, due to the significant increase in data traffic, must be capable of detecting more information, and at faster speeds, than ever before.
ePDG: One way to help off-load the macro-RAN network from high bandwidth applications such as FaceTime is the utilization of Wi-Fi and small cell networks. However, some deployment scenarios of Wi-Fi and small cells use untrusted access between the subscriber and the core network. In these types of untrusted scenarios, the ePDG becomes of significant importance to the operator it must handle high-bandwidth traffic and provide vital security for the user.
Without sufficient capability within these elements, and the scalability to cope with future demand, operators will discover very quickly that transporting video is expensive and restrictive to overall performance – ultimately negatively impacting the user experience.
The operator’s core network must be optimised to enable the most cost and time-efficient transport of video traffic. But it must also be equipped to handle volumes of that video traffic that we haven’t yet seen. It’s easy to read the various forecasts out there and dismiss them as sensationalist reports – the fact is that those forecasts are not going to be very far from the truth.
Conversational video will be a massive challenge for mobile network operators, and subscribers will be painfully unforgiving of a poor experience. And we all know they won’t blame Apple.
Author: Manish Singh, CTO, Radisys