First of all, although there are very deep similarities shared between string theory and "real world QCD" as in the AdS/CFT correspondence or in more general large N limit scenarios, it's very important to bear in mind that up to now there are no precise identifications neither between string theory and real world QCD nor on the dual resonance models as "effective theories" emergent from QCD, even more, the statement that the dual resonance models can be "effective theories" derived from QCD-like theories is in strong contradiction with the bootstrap philosophy.
With the latter in mind, I want to enummerate some formal similarities that can be seen, to some extend, as "emergent properties" in dual resonance models when this ones are viewed as qualitatively similar to "effective QCD-like theories".
Similarity one: (The low energy spectrum) The low lying spectrum of a color confined phase of any QCD-like theory is dominated by meson and baryon excitations. The obvious similarity is that the weakly coupled description of any string vacua is given in terms of open and closed strings. Open strings are analogous to mesons (flux tubes joining quarks) and a closed string is the analogue of the pomeron. After all, the spectrum of rotating open strings is what ultimately produce Regge trajectories in string theory and the $t$-channel pomeron pole can be thinked as an analogue of how closed strings appear as S-matrix poles in one-loop open string computations.
It's worth to mention that scale invariance is broken when a QCD vacuum develops the trace anomaly intuitively because in a confined phase there is a characteristic lenght scale, roughly the mass of the lightest pseudoscalar meson in natural units. String theory also has a characteristic scale in its target space, nambely the string lenght, except that this latter is a dynamical parameter (as any parameter is in the string theory).
The impresition here is, of course, that mesons and hadrons are widely more complex that just a simple bound state of its constituent valence quarks.
Similarity two:(Fermions as D-branes) Consider a QCD lagrangian with $ SU(N_{f})\times SU(N_{f})$ chiral flavor symmetry, this symmetry is spontaneuosly broken to a diagonal $SU(N_{f})$ in vacuum (details here), the resulting chiral condensate involve bilinears $\langle \psi^{*} \psi \rangle$, the Nambu pseudogoldstone mesons with the fermions $\psi$ as its constitutents can be thinked as analogous to open strings between D-branes and the fermions(quarks) $\psi$ as the analogue to D-branes. This analogy can be made precise in topological string theories were D-branes are literally fermion instertions in correlator functions (read the page 39 of this paper for details).
Similarity three: (The existence of a chiral-like perturbation theory) In the context of the similarity two it's a "miracle" that something like chiral perturbation theory exists. That's ultimately related to the fact that the inverse mass of the constitutent quarks of a pseudoscalar meson can be trated as small parameters if the mass of the aformentioned quarks is large. That's similar to the "miracle" of the integrability of the open string perturbation theory, here the very large mass of the quarks to what strings are attached to is justified because they were identified as D-branes.
