While the wireless industry has the components – especially with big data – to prevent disaster, it is reaching an Apollo 13 moment, writes Rob Chimsky
In the classic film Apollo 13, there is a dramatic moment when the NASA engineers dump a collection of parts available on the spaceship from which they must create a life saving air scrubber in a matter of hours or the entire crew will die. As happens in Hollywood and occasionally in real life, the engineers succeed in the nick of time, enabling the crew to return safely to Earth.
We in the wireless industry have hit our Apollo 13 moment. The outcome may represent less personal danger, but has major implications for the health and vitality of future wireless business. Similarly to the movie, the wireless industry already has many of the components it needs to prevent disaster with big data providing the key-enabling element. So what is this deepening crisis that will have us on the edge of our seats?
The technology advances that we have traditionally been dependent on to solve the growing appetite for wireless capacity will simply not sustain us into the future. The industry will be impacted on two main fronts. First, spectrum availability, which has always been a scarce resource, is becoming even more challenging. Exponential growth in user data consumption is gobbling up large chunks of spectrum with regulatory challenges preventing much new spectrum allocation. Secondly, technology advances such as coding schemes and modulation techniques that have led us most recently from 2G to 3G to 4G are running into practical limits. Given that these two forces will not be coming to the rescue to save the impending wireless capacity crunch, the industry must turn to current capabilities such as big data, small cells, and self organising networks that can be leveraged in order to create a sustainable solution.
Spectrum supply is running thin
The wireless industry has never been spectrum rich, but never before have we faced such a shortage of supply with minimal prospects on the horizon. The time cycles for getting spectrum to market have always been counted in years verging on decades. However, spectrum has come into the market generally with sufficient lead-time to prevent any substantial crisis.
Most recently, digital dividend or 700MHz spectrum has provided a major boost in spectrum allocation. As we look out into the future, though, there are no significant blocks of spectrum that appear to be commercially usable over the next few years. Certainly regulatory agencies in mature markets worldwide, including the Federal Communications Commission in the US, have the best of intentions in getting spectrum into the market. But the choices of future blocks all come with incumbent inhabitants who believe their interests to be as vital as those of the wireless industry. The challenge is further magnified by the fact that viable spectrum for cellular fits into a finite range that is generally below 3GHz. This situation will surely be repeated throughout the world as usage grows exponentially and spectrum allocation choices become increasingly more difficult to find.
Technology reaches practical limits
The feats of technology advancements in the wireless industry have been truly exceptional, including GSM, CDMA, UMTS, EVDO, and LTE as some of the most significant developments. These technology improvements that have allowed better throughput and capacity have essentially come from two directions. The first has been to permit larger channels to support individual traffic. For example, the move of GSM from 200 KHz channels to UMTS with 5MHz channels to LTE with even larger potential channels. In future, for many of the same reasons mentioned in terms of spectrum challenges, the ability to just carve out larger channels has practical limitations that will result in minimal benefit coming in the form of larger channels.
Secondly, technological advancements have seen a focus on areas such as coding and modulation schemes to get the most capacity out of existing channels. These advances have been spectacular in the past, but as most experts would say, we are reaching the physical limits of how much data we can stuff into an existing pipe. Previously, we had CDMA that led to significant improvements over TDMA, and then OFDMA that has created yet another leap. However, there is not another step function that appears to be on the horizon using these types of advancements.
The Apollo 13 Solution
With all this, we have reached our Apollo 13 moment. Luckily, we have components at hand that can be manipulated to create solutions with big data acting as the duct tape binding the various pieces together. So what are the elements that big data will help make use of?
1. Small Cells: Since the beginning of the wireless industry, a standard approach to gaining capacity is cell splitting. However, the next evolution of splitting will take that concept and move it from hundreds of sites to potentially thousands or hundreds of thousands of sites. We are already seeing the beginning of this with the variety of small cells going by names such as femtocells, picocells, and microcells. The concept will also incorporate the use of multiple technologies such as Wi-Fi in order to offload the macro cellular network.
2. Self Optimising Networks (SON): Integral to the effective deployment of small cells is SON technology that will allow these thousands of sites to seamlessly interact. With these thousands of sites, it is clear that the traditional macro cellular methods for implementing and optimising sites requiring significant human intervention will not go down to the small scale level. SON will be the architecture technology that permits these small sites to go in and out of the network on a very dynamic basis with minimal human intervention.
3. Cloud Networks/Virtualisation: Cloud and virtualisation will permit capacity in the core network to be dynamically allocated where it is most needed. Operators will be able to manage capacity efficiently at a large scale level, which is not possible with traditional dedicated MSOs.
Each of these elements provides an aspect of the solution but without the right direction the improvement will be sub-optimal. In order to allow the prior three technologies to reach their full potential, the solution requires the use of real-time/near-real-time data analysis, also referred to as big data analytics. The key is effectively collecting and analysing several threads of information coming from the network, users, devices, and myriad other sources. The information analysed can then be used to look for various triggers or performance levels that will then be used to automatically initiate network actions, which will optimise the capacity environment.
The big data analysis of customer geographic usage will be used to accurately pin-point where small cells should be located. Further analysis of customer patterns and network impacts will enable SON configuration to occur more accurately. Near-real-time analytics will identify changing patterns and network issues, which can then automatically initiate alternate configurations based on pre-defined parameters and instructions. In addition, big data can effectively establish trends that can be used to proactively implement network changes to avert problems before they have an opportunity to be customer impacting.
The result is that the capacity advances for the next few years will be largely dependent on these technologies – although in some cases still very nascent – and combining them in new ways. Industry, we do have a problem, but the good news is that we have the ability to leverage big data to optimise other existing tools and engineer a sustaining solution.
