What is the Google dance?

The Google Dance

There is an SEO catch-phrase called the Google dance.  Explained in simple terms, the Google dance refers to the index update of Google servers and how information hierarchies change during the update, which can take hours or even days to complete.  As the index updates proceeds, the new index must be installed on each of Google’s servers, and because of the nature of the Internet, identical searches will sometimes vary between the old index and the new one, without any apparent rhyme or reason.

In actuality, the Google dance has mostly become a thing of the past.  In the middle of 2003, Google began running continuous index updates, and even though it is assumed that Google still has to do a complete index update from time to time, there’s no longer any set period of time when such updates are likely to occur, and search results are not longer affected as noticeably as before mid-2003.  For this reason, the Google dance we are discussion in this article is a relatively historical item, and no longer of any practical use except possibly to those whose business revolves around promoting SEO/SEM, rather than those who USE SEO/SEM.  What is meant by this is that the Google dance may still be of interest to those whose profession involves promoting the use of search engine optimization, but it is not likely to be of much use to those who simply have a website that is built on SEO practices.

Estimates are that Google employs more than 10,000 servers.  This seems a reasonable number, but we must also point out that the actual number of servers used is a closely guarded Google secret, as reported by Google to the History Channel for a recent documentary on 1990′s technology trends.  While many of the servers can be identified using various IP tools, the results cannot ever be more than approximations of the total number used.  For the purposes of explaining the Google dance, 10,000 is as good a number to use as any other.

When an update begins, it is tactically impossible to update all servers at the exact same time.  Instead, Google sends the updates to its servers based on their DNS (Domain Name Service) allocations, one by one.  Some SEO experts claim that Google uses an inverted updating method, while others insist that it is straightforward, counting from the lowest numbered IP address upwards.  Either way, the data is going to be updated at one end of the spectrum before it reaches the other end.  This sequential updating is the Google index side of the dance.  The other half of the dance is related to search engine usage.

When a person performs a search query, they are not connecting to the same server index each time.  This makes it possible to receive the updated index on one search, and then immediately perform the same search, and receive the older information.  Because of the numbers of servers involved, the exact same search could cycle between new and old index files for many hours, and sometimes days, or even weeks.  Theoretically, there could even be more than two data indexes available at any given time, although most people seem to agree that there are never more than 2 Google indexes available at a single moment in time.

Google could, if they chose, rely on a central server to record and process queries, and thereby avoid conflicting data completely, but to do so would slow down the search process by a noticeable amount.  In order to maintain the speed, the Google dance is the result.  Other factors complicate the scenario a little more, but short of explaining the process of DNS is detail, the explanation given here should suffice.

Article written by SEOnotepad.com