QR2.2.4 The Scalability Problem

Berners-Lee called a scalable system one whose performance doesn’t degrade as it expands however big it gets because growth increases demand and supply in tandem, so the system can grow forever (Berners-Lee, 2000). He designed the World Wide Web to be scalable and the Internet also began this way, as every new Internet Service Provider (ISP) increases both the demand and the processing to handle it. A scalable network has to distribute control but when the Internet began, pundits expected its lack of central control to result in chaos. It didn’t collapse and the reason turned out to be because it had no central control. Computer science discovered that an infinity anywhere in a centralized network crashes it but distributed networks carry on despite a local crash – they degrade but don’t collapse. Our brain neural network distributes control for the same reasons (Whitworth, 2008).

The performance of our space hasn’t changed much over time, even after expanding for billions of years, so if space isn’t nothing, it must be a scalable system. If our virtual space expands like the Internet, then adding nodes must increase both supply and demand, suggesting that space has local limits. The evidence agrees as:

“…recent observations favor cosmological models in which there are fundamental upper bounds on both the information content and information processing rate.” (Paul Davies, 2004) p13.

We call the upper limit of what space can hold a black hole. It is in effect the bandwidth of space.

In general, space as a scalable network suggests distributed control.

In contrast, Cartesian coordinates require:

1. A predefined maximum size: Cartesian coordinate memory allocation requires a predefined size, so a point stored as (2,9,8) in a 9-unit cube must be stored as (002,009,008) in a 999-unit cube, and so takes up more memory.

2. A zero-point origin: A (0,0,0) point that is the absolute center of space.

A Cartesian space needs a predefined maximum size but our space has expanded for billions of years with no end in sight. If our space was Cartesian, its maximum size would have to be defined before the first event, to avoid a Y2K problem. Our Y2K problem arose because old computers stored years as two digits to save memory, so 1949 was stored as “49”. The problem was that the year 2000 would be stored as “00” that was used for the year 1900. Changing all our databases to four-digit years meant that any programs accessing them might crash if not modified. That our space is still expanding without upgrade suggests that it can’t be a Cartesian virtual space.

A Cartesian space also requires an origin point from which to expand. Since Hubble showed that every star and galaxy is receding from us, a Cartesian space implies that our Earth is that origin! Planet earth only began recently, so it can’t be the Universe’s origin. Our space is expanding with no absolute center so it can’t be Cartesian.

In general, Cartesian coordinates work for small spaces but not for an expanding space as ours is.