“To be, or not to be? That is the question”
   ― William Shakespeare, Hamlet

Reflection on Data Fidelity

Is data communication or more specific, computer networking, science or art? Is the world analog or digital?

Ever since computer started taking charge of our life, everything started the transformation of digitization, sooner or later. It makes everything easier to quantify, measure, compute and transport and is truly the foundation for the information revolution.

Many years of engineering training and over 20 years of professional practice in data communication industry, it has become my second nature to think and view things in 0s and 1s, black or white, right or wrong. Everything becomes easier and manageable this way and you can really feel the power to change the world. As we immersing deeper and deeper into this digital world we help created, I realized one day that something is not right, I started to get confused about the truth vs lie, fact vs fiction, and more importantly, purpose vs meaning.

Rethinking what is the purpose of data communication, is it really just sending data from A to B accurately, efficiently and in an timely manner?

I would say that’s mostly (will elaborate more later) true among digital computers, but surely not totally true for human beings.

For human beings, digitalization is never by nature. Our brains doesn’t work like digital computer and our perception system is NOT in binary. Our sensing system works continuously within its functional spectrum, our intelligence system takes infinite inputs and everyone evolves into unique beautiful beings. One of my hobby is working on my HiFi system and one of lessons I learnt from it is that digital processing and amplification does create lots of fun and flexibility, it NEVER sounds as nature as an analog system. 0s and 1s does reduces ambiguity within the player/receiver/amp but it’s not the fidelity we human being want as listener. From this example, clearly we can see the digital data communication system we use today has three fundamental functions:

  1. A/D and D/A transformation at input and output to inject the digital system into a naturally analog world
  2. Digital processing, such as encapsulation, multiplexing, encoding etc
  3. Recording or transportation

Traditionally, data communication industry focuses on function 2 and 3 above and leaves function 1 to human-machine interface devices like computer and mobile devices. Taking the HiFi case as an example, function 1 decides the fidelity of music we hear out of the system no matter how well we do in function 2 and 3. As a matter of fact, regarding to fidelity, the best we can do in function 2 and 3 is to reduce the degradation to ZERO. Yeah, that is pathetic but it’s exactly the problem we have in this industry. For too long we have buried our head in sand building pipes, eventually we lost our sight for the purpose of communication. We struggle on pushing towards Shannon’s Theorem limit, building more and more, fatter and fatter data pipes, from deep sea to outer space but still, we see no future to meet the mounting demands.

It’s time to drop the wrench and rethink hard on the purpose for communication, which fidelity we should be focusing on now. Here are a few questions worth asking ourselves:

  1. Garbage in, garbage out - should we care A LOT on what to communicate, aka the input, in stead of just making sure transport “garbages” in high fidelity?
  2. Do we communicate the **RIGHT” information, no more and no less, over the communication system? Instead of trying to beat the Shannon limit, shouldn’t we change the battle field rethinking what should be the “Information” we feed into the pipes as a better approach?
  3. Can we make the RIGHT communication choice when communication system is peripheralized from the rest of information processing systems, namely compute and storage?

Think of an interactive Holographic Telepresence scheme, which transports only 100s Mbps meta data in stead of 100-1000s Gbps encoded 3D data, that is 3-4 orders of magnitude reduction in both bandwidth and latency demand to the communication system, making it a much achievable goal.

This is a challenging time for networking, but it’s also a perfect time for breaking away: Welcome to an era of converged compute, storage and networking, friend!