patternMinor
How come Google search never gets slower? Doesn't it have to go through the ever increasing amount of data on the internet and then rank?
Viewed 0 times
googlecomerankamountdoesnthenhoweverdatathe
Problem
I heard Google creates indexes of pages. Even then, the number of indexes goes with time. Why wouldn't this slow down the search?
Solution
The google indexing and search is run in parallel using a map-reduce algorithm.
As the number of pages to index expands, so does the number of computers that google uses for the task.
Currently Google handles about 3.5 billion searches per day.
A single search uses 1 000 computers in parallel (each computer holds $\frac{1}{1000}$ of the total Google index) and takes on average 0.2 seconds.
Servers available for search are always in read-only mode. When the search index gets updated those servers are temporary unavailable for search.
Given the current amount of floor-space in its data centres Google has a maximum of about 1.7 million servers.
Obviously most of these are devoted to something other than search.
As the number of pages to index expands, so does the number of computers that google uses for the task.
Currently Google handles about 3.5 billion searches per day.
A single search uses 1 000 computers in parallel (each computer holds $\frac{1}{1000}$ of the total Google index) and takes on average 0.2 seconds.
Servers available for search are always in read-only mode. When the search index gets updated those servers are temporary unavailable for search.
Given the current amount of floor-space in its data centres Google has a maximum of about 1.7 million servers.
Obviously most of these are devoted to something other than search.
Context
StackExchange Computer Science Q#50653, answer score: 3
Revisions (0)
No revisions yet.