Can bandwidth growth meet exploding demand?

Moore’s Law Doesn’t Apply to Bandwidth

The internet is arguably humankind’s greatest technological innovation. It has transformed business, government and our personal lives, eroding political borders, creating alliances and new identities for people separated by previously unbridgeable geographic distances and socio-cultural differences. Unfortunately, while it has increased opportunity and positive connection for many, it also has ‘retribalized’ others, even catalyzing terroristic acts among some reactionary groups and easily manipulated ‘lone wolves.’

Moore’s law and computational power

Moore’s Law states that computer power doubles every two years. Most take for granted that this dynamic will continue into the far future. Taken to its limit, something like the ‘singularity’ predicted by Ray Kurzweil could trigger the development of a superintelligence that would trigger inconceivable changes to life, as we know it.

Does Moore’s law apply to increased data transmission speed?

However, Moore’s Law doesn’t apply to data transmission rates/bandwidth. Making that easy assumption threatens to distract IT network engineers from developing strategies to handle tomorrow’s insatiable demand for greater connectivity and speed. Save for the possibility of an unimaginable breakthrough, networks are approaching a limit. More specifically, scientists have described a ‘nonlinear Shannon Limit’ on the quantity of data that can be transmitted through any channel, just as there is a limit to the speed of light.

Why is this common assumption wrong?

  • Speed is determined by bandwidth AND latency. Latency is a term that describes how equipment slows the transmission speed of signals This even includes delays caused by direct transmission from satellite to receiver. Despite recent reductions in latency, there’s no way to move beyond inherent constraints on speed, even for speed-of-light fiber optics. (Fiber optic cables require amplifiers that slow things down).
  • Networks also have ‘schedulability’ limits, i.e., data have to be dissected and rearranged for better outcomes, which also slows things down. This is another aspect of latency, not bandwidth.
  • Information ‘packets’ cue up along transmission links and create competition (aka ‘contention’) for delivery.
  • When bandwidth increases, demand automatically increases to fill capacity. This can cause regression in performance as excessive demand undermines the stability (aka “stationarity”) of transmission.

Bottom Line–While increases in computational power are governed by Moore’s Law, data transmission speed and capacity have limits that are likely not only to exceed demand but also work against reductions in cost.