The Covid responses across the world, across governments have been very diverse, modulated by the local geography and other functions like population density, demographics and the behavior of a given population. The similarities between International Covid responses and the modern wireless RAN design have been explored in this blog post.
Since the Covid pandemic outbreak earlier in the year, it is interesting to note the responses of various governments all over the world. While one would have expected isomorphic governance models like democracies to respond similarly, there have been variations across countries all over the world. The framework though has been the same namely lockdown-unlock loops the “time constants” of these loops and the manned of the lockdown has been different. And rightly so. WHO has provided guidelines but not diktats. The responses also have altered over the months as the virus statistically has gone from being an “unknown unknown” to a “somewhat known unknown”. A lot of statistics about the fatalities and the medical responses on various populations has become available since the outbreak of the pandemic.
Most democracies have responded with various shades of lockdowns. In India there was a very stringent lockdown in the last week of March that lasted for more than a month. Though India is a federal republic, the lockdown decision was almost a diktat from the central federal government. The unlocking though was a federally decentralized decision with the centre deciding a guideline and letting the states decide the process of unlocking. Though there were issues with the migrant crisis which could have been handled much better, the government has been responding reasonably quickly based on the data available. It does make sense in hindsight (the process and not the decision!). The virus has followed a predictable Pareto principle (80-20 or 90-10 rule) with almost 90% of the cases concentrated in less than 10% of the geography.
The federal government US in contrast let the states decide when to lockdown and unlock right from the beginning. Sweden was the extreme end of the other spectrum where it was left as a discretionary decision to the citizens. Other democracies have been between the two extremes. The process to me is more important than the outcome. I am not arguing that one decision was better than the other. The statistics speak for themselves.
The responses clearly have been the way it should have been. Very diverse modulated by local geographic compulsions and other functions like population density, demographics and the behavior of a given population. Imagine if the WHO (it can’t but as a thought experiment) had issued a diktat of a single response to the situation all over the world.
What’s all this got to do with wireless RAN designs? A lot and hopefully at the end you might get my point. The real world an” y complex system is fundamentally non-linear and statistical. We tend to linearize “reality to comprehend it better and apply reductionist tools. That makes the math easy and gets you published in a scientific journal! A complex system needs to be “loopy” and in such systems the best way to reach equilibrium is to apply some kind of recursive stochastic reasoning based on Bayes.
Let’s first understand the process of standards making. Wireless networks from 2g to 5g need protocols at the “physical” and “logical” levels for devices to interoperate and work seamlessly. This is why your iPhones and Andrioid phones can communicate with a wide variety of wireless infrastructure deployed by operators all round the world. These standards are defined by 3gpp an internal organization made up of member companies very much like the UN or the WHO. And guess what all the politics plays out in the UN also gets played out in the standards making process. Technically 3gpp is an open standard just as UN is but is dominated by a few players with financial and market clout. So all the geo-political issues now get projected and carried over to these “non political” bodies as well. A standard is important as it provides economies of scale for equipment manufacturers and is one the reasons why we get cheaper semiconductors and therefore phones.
The standard making process is a complex one and a refresh cycle and a transition between subsequent generations generally takes between 5-10 years. From a modem/radio perspective these standards tend to be “static” and are designed for worst case conditions.
Radio environments are extremely dynamic, changing over time and space. A given cell site in Mumbai versus say in a suburban European city or a village in India can experience very different conditions due to terrain, network traffic conditions and weather conditions in some cases. So a “static” unchanging worst case design cannot cater to these almost infinite space of conditions. Such a design results in a sub-optimal network wasting precious spectral resources as spectrum is a very expensive commodity.
So just like WHO advisory to Covid, 3gpp can best serve networks by issuing a guideline framework under which every cell site can take its own decentralized decisions based on the current radio conditions and cost functions that need to be locally optimized. At Saankhya we are building a set of technologies and IPR that will allow us to “map” the radio environment and build an AI-RAN that self optimizes based on conditions. We call this the AI-RAN and the process of generating maps “Radio mapping”. All this is now possible as a convergence of AI/ML, Virtualization and cloud computing that have now matured as technologies
In conclusion, complex non-linear dynamical systems like Covid responses and optimal wireless RAN designs mandate the use of a decentralized decision making based on local stimuli/response.