A Scalable Parallel Approach for Peptide Identification from Large-Scale Mass Spectrometry Data