Bacterial identifications are now usually done using their 16S rRNA gene sequences. The reasons for using the 16S rRNA gene are twofold:
  First, the 16S rRNA gene is sufficiently distinct from its eukaryotic, archaeal, mitochondrial and chloroplastic homologs, so that each can be amplified separetely. It is however enough conserved so that almost universal bacterial amplification can be performed using specific primers.
  Second, the 16S rRNA gene sequence is generally species specific: the microbiologist or clinical researcher can not only determine the diversity within a bacterial community after alignements and classification (or phylogeny), but the sequence itself allows generally for species determination (with a few counter examples).

There are two usual applications:
  One wants to study the microbial community within a given environmental or clinical sample. The wast majority of environmental bacteria cannot be cultured in the laboratory while but their DNA can be easily extracted. From this DNA, identification using the 16S rRNA gene sequences is usually performed, after cloning the PCR products and random sequencing of tens to thousands of clones.
  A strain has been isolated, and one wishes to know if it is a well known species or a new one.

In both cases, the common approach is to blast the new sequence(s), align it to the most similar sequences found, and do a classification or a phylogenetic analysis in order to decide about identification. Even when doing biodiversity analyses, it is generally best not only to find out which are the most similar sequences present in the public databases, but also which well known cultured species is closest (in order to have an idea of which kind of biochemical process they can contribute to, for example).