The apparent analogy between the self-diffusion of linear oligomers in monodisperse systems, 2 up to 32 monomers, and their tracer diffusion in an entangled polymer matrix of length 256 is investigated by molecular dynamics simulations at constant pressure. Oligomers and polymers are represented by the same coarse-grained (bead-spring) model. An analysis based on the Rouse model is presented. The scaling relationship of the self-diffusion coefficient D with the chain length N written as D proportional to N-alpha is analyzed for a wide range of temperatures down to the glass transition temperature T-g. Near T-g, the heterogeneous dynamics is explored by the self-part of the van Hove distribution function and various non-Gaussian parameters. For the self-diffusion in a monodisperse system a scaling exponent alpha(T)>1 depending on temperature is found, whereas for the tracer diffusion in an entangled matrix alpha=1 is obtained at all temperatures, regardless of the oligomer length. The different scaling behavior between both systems is explained by a different monomer mobility, which depends on chain length for monodisperse systems, but is constant for all tracers in the polymer matrix.