Optimizing a service provider network

Broadband access and broadband services will be the main driver of revenue growth for mobile service providers over the next five years. However, there is currently a significant strain being placed on wireless networks. With smartphone and tablet adoption on the rise and user demand for mobile video continuing to explode, service providers are increasingly challenged to manage expanding networks.



Broadband access and broadband services will be the main driver of revenue growth for mobile service providers over the next five years. However, there is currently a significant strain being placed on wireless networks. With smartphone and tablet adoption on the rise and user demand for mobile video continuing to explode, service providers are increasingly challenged to manage expanding networks.


Optimizing a provider network is a complex task. We’ve all heard about poor network performance because of “clogged” networks.  And it is not surprising given the growth in mobile data traffic as shown in the chart below.  Service providers need to enable effective transport of VoIP, data, and video traffic in 3G and 4G/LTE mobile and next-generation switching networks so that subscribers have a good experience using these networks.  



Click image to enlarge


There are various ways to optimize a network as indicated below. Upgrading the network, deploying new capacity, using optimization equipment and deploying new technology are all options that involve spending money. There are huge amounts of money being spent on upgrading 3G networks to faster speeds, LTE technology and also building out WiFi and Femtocell networks.



Click image to enlarge


To make the problem even harder to solve, providers need to do all of this in a significantly reduced CAPEX and OPEX world. This is because even though the subscriber numbers and subscriber usage of these networks is increasing, the revenue increase from these subscribers is not proportionate.  According to the Infonetics graph below, in 2011 there was about $778 billion of mobile revenue from voice and data, with about a 6% growth from 2010 and even less percentage growth in the coming years expected.  Broadband growth, on the other hand, is at an over 25% CAGR in the years from 2010 to 2015, according to a December 15, 2011 repot from Infonetics titled “2G/3G/4G Services and Subscribers: Voice, SMS/MMS, and Broadband.” 



Click image to enlarge


Why the difference between the huge percentage growth in data traffic and the relatively tepid growth in revenues? Even data revenue CAGR doesn’t match the huge mobile data growth.  This is because of monthly data plans and the like. The chart below from Wireless Intelligence shows the Average Revenue per user decrease.  The red line shows the weighted average ARPU of mobile broadband operators with both voice and data networks, and the grey line shows ARPU of the non-mobile broadband operators with only voice networks.  Both are declining, but the mobile broadband operators with both voice and data networks show less of a decline. Either way, there is more traffic, there is less money to work with, and therefore there is a reduced CAPEX and OPEX world creating additional challenges for operators.



Click image to enlarge


In the Infonetics graphic included above you can see that in 2011 there is over $100 billion of data service revenue. This is growing at almost 25% per year and becomes almost 30% of the overall service provider revenue by 2015. With regards to video specifically, it’s more difficult since much video is via OTT and that is likely not monetized.  If video becomes two-thirds of the data services by 2016 as forecasted by Cisco, then that’s one way to look at it.  Also, ABI Research has done video telephony research and in a July 2010 report titled “Mobile Video Services” they estimate about $1 billion or more of monetized video services in the 2012/2013 timeframe.


Video is clearly where the opportunity lies for service providers, but how can they ensure their networks can handle it?


Putting in specific optimization equipment is a good option and mobile backhaul optimization is a growth area for that very reason.  Mobile backhaul is the mechanism by which traffic goes from the cell tower to the core of the network. Ovum estimates that by 2015 mobile operator revenue will be greater than $1 trillion dollars and that service providers will be spending over $180 billion on mobile backhaul transport costs with the backhaul equipment being an $8.7 billion market, according to their 2012 Wireless Backhaul Forecast. The equipment will take many forms – optimization from caching mechanisms, optimization from compression mechanisms, optimization from content and device awareness mechanisms.



Click image to enlarge


Optimization can also be achieved if service providers can control the original content travelling through their networks. By ensuring the network is device aware, content aware, and network aware, service providers can ensure that the most optimized video stream is sent over the network to the end device. 


Probably the best example of a network being device aware is your browser.  Many websites render the website differently to a computer or laptop than to a mobile device. The website to the mobile device can be a mini version of the laptop looking website, or can be a different version, tailored to deal with what the mobile devices are doing on the website.  Essentially, the website for the mobile device wouldn’t have all the information and what it had would be easy to navigate through.


Regarding video specifically, there would be examples of the network (or specifically the CDN or service provider) knowing your device can’t handle HD video format.  If it knew that, then the CDN wouldn’t send over an HD video format (which takes more bandwidth) since it wouldn’t make a difference.  It would send over the video format that was best suited to your device. In other words, only send over the format that the subscriber can utilize – the subscriber gets the same experience, and bandwidth is saved or more accurately optimized on the network.


Also, it’s been widely reported that many videos viewed online aren’t viewed fully to the end of the video. So if the video is cached in the network, then whatever is cached and not viewed is bandwidth that is wasted.   An example of network awareness would be whether the viewer is watching the video via WiFi, 3G, or 4G. If it is on 3G, then the provider should only send the video format that is compatible with the 3G network.  Sending a full HD video format on a busy 3G network will simply result in a poor user experience.


In order to monetize on mobile video, service providers also must conduct Quality of Experience (QoE) analysis. Understanding what the subscriber is seeing versus what was originally sent is a key tool for service providers.  For example, if a subscriber has a poor mobile video experience, it could be because of factors such as poor transcoding, poor caching, actual bandwidth issues or the wrong format being sent to a device. If a service provider or content delivery network knows exactly where the quality degraded, then service level agreements can be measured and the subscriber experience can be known instead of guessed. This ultimately allows service providers to monetize mobile video through tiered services or premium pricing models. 


Many of us have experienced real life QoE analysis in the form of answering a survey after viewing a video on-line or via a mobile device.  However, the industry is getting more sophisticated now.  There is network infrastructure and software now that can help measure the quality of what the subscriber is seeing. 


For instance, I spoke above about different video formats to different devices.  If, say, a sports game is being distributed, then it might need to be sent out in 12 different formats (HD format, non-HD format, small screen, tablet size screen, Android device, iOS device, etc.).  Maybe it’s even more than 12 formats, as you can easily see the different combinations would add up. Regardless, it would be nice to know if the transcoded/transrated/transsized video had near the same quality as the original before it was sent out. And then if the quality is not good enough compared to the original, it could be redone.     


Another example would be measuring the original with what is being sent, somewhere at a node in the network.  CDNs and service providers have SLAs.  Measuring and help understanding where an issue is would help enforce SLAs as an example.


Ultimately though, instead of filling out a form about your experience, the “network” might already know.  And if the experience was not good, which could be from a variety of factors, then whoever you paid money to watch what you watched could be proactive with you about it.  And that would be good customer service and you’d be apt to try it again. 


There is no question that mobile video is not going away but only growing in demand, especially with smartphone devices surpassing feature phones this year. This presents an enormous opportunity for service providers to capitalize on mobile video services, particularly if they can offer high quality video services. The first step for providers will be to overcome the current obstacles: enabling their networks to handle the demand and then monetizing their services.