The future of access networks
by Maarten Ectors on 16 September 2015
IoT, Docker, Big Data, Cloud, mobile apps, Video on Demand, etc. It seems like innovation is accelerating but without a roadmap, business plan or strategy. Google and co. seem to have their eye set on taking over the telecom business. Google’s global fiber network, SDN, Fiber, Fi, OnHub, Android, Hangout, Brillo, White Spaces Spectrum Database, Loon, etc. are all targeted to dominate a certain telecom segment. However what will happen when they start to overlap?
As a telecom operator, every year more data needs to be supported, people spend less on data plans, people want to pay less for calls & SMS or even hardly use them. Broadband and mobile plans have become commodity. Over the top revenue is virtually nonexisting. IoT promises a data tsunami but not a revenue tsunami.
Current suppliers are fast when it comes to selling new equipment. But every generation seems the same as the last. None really addresses the above issues. Telecom operators have come to depend on a group of suppliers that prefer a politically correct hard to implement standard over delivering easy solutions. Patents, obscure standards, secret interfaces, etc. is their weapon against becoming obsolete. To be fair they are trying to win RFPs that have never given points for having an easy, beautiful or innovative solution.
The telecom business has been like the unbreakable Nokia 3310: robust, fit for purpose, universally available and standards driven. Unfortunately we moved into an era where agility, hyper innovation and giving customers a wealth of choice is higher rated. Apple and Google will write the future of the telecom industry unless somebody really starts challenging them.
How to fix the biggest telecom industry problems?
Let’s start with CPEs. Cheap onpremise equipment that implements basic functions like NAT, DHCP, WiFi, etc. are large cost centres. Equipment is expensive to purchase, maintain and support. Too many calls to the call centre about how to open a certain port. The future of CPEs are microservers. Small cheap supercomputers in each home or business that can be personalised with apps and app stores. Why would customers want apps on CPEs?
Let’s take a look at one consumer use case: parental control. The current version of parental control either consists of manually going into the CPE admin page and introducing URLs that are blacklisted. Alternatively there is a network approach whereby all people in the home are banned from certain content. The first one is not a technical option consumers are comfortable with. The second one puts the operator at risk of being seen as either Big Brother or making embarrassing mistakes.
Via an app store worried parents can pick different solutions for different children. The 3 year old can connect to a fully managed Disney experience in order to make sure they are busy 1 hour every day. The 7 year old can only play online 20 minutes, watch videos 30 minutes after 19:00 and can always go to Wikipedia. The 16 year old can navigate without filters but if 5 times a bulimia site is visited in 8 hours, parents will get a list of experts to speak to. The operator will get a revenue share for all apps that are purchased. Hence all of a sudden the CPE becomes a revenue generator. The good thing is that the operator does not need to go and create their own ecosystem of apps. They can have a fully branded and governed app store that piggybacks on an existing ecosystem of apps with the possibility to launch exclusive apps.
Since the CPE is a microserver some of the typical network functions like NAT, DHCP, etc. can be moved to the central office where a minicloud can provision and manage them as well as control other aspects of the software defined network. Deploying an SDN agent is just an app.
The CPE has WiFi, Bluetooth, potentially Zigbee, and as such it can do building automation & security. IP cameras can talk to it. So can the HVAC and the alarm system. All of a sudden the CPE becomes the IoT hub of any business or home. After installing 5 apps and interconnecting the heating or the lights, no consumer or business will want to change broadband provider because it is just too cumbersome. Zero churn.
The costs of the CPE are partially or completely subsidized by the hardware vendor and are covered by a revenue share the hardware vendor gets. This way costs will be equal or lower than existing less powerful solutions and even in the future CPEs might be fully subsidized just like mobiles are. This time however by suppliers.
Let’s focus on the next network element: the DSLAM or the mobile base station for that matter. App enabling the DSLAM will allow other types of use cases. By including extra storage and compute capabilities on the DSLAM, mobile base station or potentially the central office, these can be rented by the hour or GB. What would that bring? Having a spot market would allow companies to bid for them each hour. This will allow a Netflix, World of Warcraft or AC Nielsen to compete for the same compute and storage because either the latest 4K hit series is cached on the curb, game virtualization or anonymized market analysis is done. The end result is substantial over the top revenue. Server side components of a device/phone solution could be ran as well. A parental control could consist of a proxy that runs on the DSLAM or the mobile base station and follows personalized rules.
Of course you want your phone, tablet and PC to all interact with this new app world of possibilities. But why have all this complexity. Why not have one device that does it all? A convergent device that is a mobile in your pocket but at work or at home turns itself into a full PC that can use any large screen and Bluetooth keyboard you can get your hands on. Mobile apps will have their bigger brother equivalent hence they share the same data but are optimised for each screen size.
Does this all sound like science fiction? Well it isn’t because the open source snappy Ubuntu Core can be used to app enable any type of device and the open source Juju can be used to setup and control any mini clouds as well as provision, integrate and scale solutions in minutes close to the edges of your network. The Ubuntu Phone will be your convergent device that is both your mobile and your PC, with apps scaling up based on your needs.
Do you already have regular providers offering you apps for CPEs? Perfect. However how easy is it for developers to create an app? You can’t measure easiness in an RFP. Can your regular supplier explain in under two minutes on Youtube how a developer creates an app, uploads it to the store and puts it on any type of device that can run Ubuntu Core, e.g. drones, robots, fridges, CPEs, IoT gateways, VoIP gateways, top of the rack switches, etc.? We can.