What does robustness means in SVM algorithm?



I am currently studying about the SVM algorithm and while studying it I came across that SVM is the algorithm which find the best suitable line for classification of given classes and it also maximize the robustness but I am unable to understand what is the robustness and why it always tries to maximize the robustness.

Here in this figure the middle line is the one which separates two classes and also increase the robustness of SVM algorithm. I wantto know how middle line increases the robustness of SVM algorthim



The margin is higher with the vertical line, the other lines nearly touch a point therefore low margin and difficult separation between your two classes.