How much faster is a Google server than a normal PC?

Besides the fact (as others have mentioned) that Google (and other major sites) do not use ONE server but millions of them in multiple giant computing centers, they also use another trick to respond faster: they store a lot of the common elements of the web pages in “edge cache” that is near your home.

It works like this:

Look at the web page you are viewing right now. There are parts that are very specific responses to your input, but many things are pretty generic and *everyone* sees them. Some are even quite static - icons and logos that don’t change for months or even years.

When you access a host system, your inquiry may go a long way over the internet to get to the host, and the host responds over that same path. That takes time. We might measure it in fractions of a second, but it still adds up.

So, a long time ago someone got clever and figured out that they could save time (and bandwidth) if the content that many people saw was stored near their location instead of at the host site. So they pre-positioned things like graphics that require a big file at ISPs who agreed to host their content. This meant that the time it takes for you to get their entire web page is much quicker.

For some, the speed of response was critical to their business. If you do a search, for example, you want the answers to come as quickly as possible.

These days, it is common technique used by all the big web sites. Tens of thousands of internet service locations support “edge cache”. Back in the late 1990’s, my job was to develop the procedures to implement 5,000 new locations per year, using satellites to broadcast the common content all over the world. We gave each participating ISP $30,000 in satellite equipment, which reduced their connection fees.