This uses the taxonomic definition of race, which http://www.nature.com/ng/journal/v36/n11s/full/ng1455.html says "Although 'race' and subspecies are usually treated as equivalent, some zoological taxonomists reserve the word 'race' for local breeding populations, with subspecies being geographical collections of populations that are similar or the same in the defining traits."
That same article is careful to point out that the taxonomical definition is different from the sociopolitical, "socially constructed" definitions used by humans when talking about themselves.
Concerning dogs, https://en.wikipedia.org/wiki/Dog#Taxonomy points out that it's unresolved if domesticated animals should be list as a subspecies, so that dogs are "Canis lupus familiaris", or if they are their own species, as in "Canis familiaris".
In either case, dog breeds are "artificial variants created by ... selective breeding", from http://vertebrates.si.edu/msw/mswCFApp/msw/taxon_browser.cfm... . The term "variety" has no formal definition in mammalian taxonomy. See https://en.wikipedia.org/wiki/Variety_(botany)#Other_nomencl... .
Dogs breeds are not generally considered as different species despite it being impossible for many to interbreed (think chihuahuas and Great Danes). Isn't biology wonderful :)
Are different dog breeds the equivalent of human races, for example?