There sometimes is a need to offer more than one page on a website where the pages essentially contain the same content. Duplicate content could be seen as unnatural by search engines and this could lead to penalties or the pages could compete against each other and dilute the URL properties of each.
Google introduced Canonical URL or Canonical link identification tag to ensure that site owners have more control over the URL returned in search results and to make sure that properties like link popularity are consolidated to the preferred URL. This tag is supported by Google, Bing, Yahoo and Ask.
The purpose of the Canonical link is similar to that of the 301 redirect which is to transfer links to the canonical page or preferred URL.
The approach involves inserting a tag into the head section of the pages that have duplicate content to specify the preferred (URL) version to ensure that that the duplicates refer to the canonical URL and any links or indexing properties are be transferred to this preferred URL.
In addition to being used for duplicate content, the rel canonical link is also used when to indicate a page that is a super-set of the component pages. The important thing to remember is that search engines only index the content from the canonical version and the content of the component pages is not indexed. However, links and indexing properties are transferred from the component pages to the canonical version.
Note: Google allows canonical links that point to completely different domains, however, Bing ignores canonical links that point to a domain that is different from the parent domain.