The Small Cell World Summit just concluded, so I would like to summarize my thoughts and impressions from this event. For a little perspective, the conference is sponsored by the Small Cell Forum, which was called the Femto Forum before it rebranded in February 2012. In alignment with this new vision, the 2013 edition of this conference featured for the first time a stream focused on backhaul for small cells. In addition, the scope was expanded to include many ecosystem players: silicon players, DAS OEMs, satellite communication service and equipment providers, protocol stack vendors, traffic localization and network optimization tool developers, timing and synchronization vendors, and engineering service companies, etc. Operators from all continents were on hand to share their experience. Equipment vendors included also In short, it was a microcosm of Mobile World Congress with a focus on wireless infrastructure, without the large peripheral crowd.
While there is much focus on outdoor small cells, as highlighted by the backhaul stream, the confluence of DAS and Wi-Fi equipment show that small cells are much more than a single category of equipment and technology. Instead, small cells are any RF transceiver node in close proximity to the mobile user operating in licensed or unlicensed band and using a number of radio access technologies (3G, LTE, Wi-Fi).
The wireless network architecture has been ‘flattened’ to an IP network through LTE. Seamless Wi-Fi will greatly simplify access of mobile users to Wi-Fi networks. Users in the future may not be able to tell if they are being served by a licensed band access network or a Wi-Fi network: concurrent traffic with different quality of service levels can be routed on different networks.
This creates major challenge: how to manage and optimize a network made up of different wireless layers and technologies? How to minimize and manage interference between the different layers? How to route traffic in a most effective way as to maximize network efficiency and capacity? There is a large number of parameters to control. With the network being layered, RF transceivers are much greatly liable to affect each other’s performance than in the traditional honey-comb cellular architecture, potentially leading to a destructive cascade of events.
The solution clearly lies in having the capacity to automate a process that currently requires human intervention: in other words ‘self-organizing networks’ (SON). But SON has been a greatly misused description, tagged to many type of activities. In the sense of heterogeneous networks, SON ought to be the ability of the network to automatically configure its parameters in a manner that optimizes performance at any given moment in time. This requires a new, expanded framework for network monitoring and performance management that goes beyond the capabilities of current OSS or network management systems provide.
The SON framework would incorporate big data techniques to make sense of the diverse and large amount of data that various network elements spew out. Machine learning techniques are used to predict the performance of the network given changes in certain parameters. Together, such analytics are the key to small cell deployments: it will define the network of the future.
This is a matter that has not been lost to operators who cite the lack of such a system as one of the main challenges to large scale small cell deployments. Companies small and large are currently working towards solving this challenge, ours included. In short, the smarts in the network is what will matter most.
 Xona Partners provides services centered on big data, machine learning and cloud services which we couple with our deep expertise in wireless networks both in the core and the RF interface.