There is little doubt that cytotoxic T lymphocytes (CTLs) can kill tumor cells in-vivo. However, most CTLinducing immunization protocols examined so far in cancer patients have yielded only limited clinical benefits, underscoring the urge to improve current approaches for the effective induction of tumor-reactive CTLs. The tumor side of the immunological frontline is armed with large masses, high mutability and an arsenal of immune evasion and suppression mechanisms. Accordingly, the confronting CTLs should come in large numbers, recognize an assortment of MHC class I (MHC-I) bound tumor-associated peptides and be brought into action under effective immunostimulatory conditions. Naive CTLs are activated to become effector cells in secondary lymphoid organs, following their productive encounter with MHC-I-bound peptides at the surface of dendritic cells (DCs). Therefore, many cancer vaccines under development focus on the optimization of peptide presentation by DCs at this critical stage. The elucidation of discrete steps and the subsequent identification of inherent bottlenecks in the MHC-I antigen presentation pathway have fueled elaborate efforts to enhance vaccine efficacy by the rational targeting of proteins or peptides, formulated into these vaccines, to this pathway. Protein- and gene-based strategies are accordingly devised to deliver tumor-associated peptides to selected cellular compartments, which are essential for the generation of functional CTL ligands. Many of these strategies target the conventional, endogenous route, while others harness the unique pathways that enable DCs to present exogenous antigens, known as cross-presentation. Here we dissect the intricate machinery that produces CTL ligands and examine how knowledge-based cancer vaccines can target the sequence of workstations, biochemical utensils and molecular intermediates comprising this production line.