Metcalfe’s Law, Moore’s Law and Google

My Take this week in SD Times News on Thursday discussed a fascinating presentation from Jonathan Rosenberg (pictured), senior VP for product management at Google. In the column, I made passing reference to Metcalfe’s Law and Moore’s Law.

Since I didn’t describe these two laws, and referred to them in adjacent paragraphs, some readers thought that one reference was a typo. It wasn’t. However, let’s use the opportunity to briefly describe these two laws.

Metcalfe’s Law, as proposed by Ethernet inventor Bob Metcalfe, says that the value of a telecommunications network is proportional to the square of the users of the system – that is, the number of potential connections between the users.

Think about fax machines, or e-mail: The more people who use it, the more useful the system is. The same concept also applies to information sources: The more books you have in a library, or the more Web pages are indexed by a search engine, the more popular it is, the more likely people will want to use it (because they’ll more likely to find what they want), and the more people will want to add more stuff to it (because it has more users).

Note that some experts agree with the principle of Metcalfe’s Law, but argue that the correct ratio is n log n, not n squared. While that intuitively seems more accurate for very large networks, I don’t have a strong opinion one way or the other.

Moore’s Law, based on observations by Intel co-founder Gordon Moore, is widely quoted saying that, for a fixed cost, the number of transistors on an integrated circuit doubles every 12 or 18 months. However, Moore himself later that he meant that the number doubles every 24 months.

For my purposes here (and in my Take), the important concept is that technology growth is exponential in many areas of computing technology, including raw CPU power, memory, storage, I/O bandwidth and network bandwidth. Or, to look at it another way, the cost of CPU power, memory, storage, I/O bandwidth and network bandwidth is decreasing at an exponential rate.

When you combine these laws, Google’s business model depends on two predictions being true for many years:

* In a Metcalfe’s Law sort of way, the amount of information that Google stores and delivers will continue to grow – and the more information Google has, the more users it will attract. The more users it has, the more advertisers it will attract. The more money advertisers can make, the more likely they’ll be to put more content there. That will attract more users, and so-on.

* In a Moore’s Law sort of way, the march of technology will make Google’s server farms faster and able to perform more complex processing, and store more content and it will make end users’ computers more powerful and it will increase the width of pipes that deliver content from Google’s service farms to end users.

That, in turn, will enable the processing, storage and delivery of yet more content, making the cycle ever more virtuous.