Can junk DNA be seen as "potentially useful junkyard parts" that some random mutation might re-activate into a gene or part of a gene? Is it actually handy to have these around to allow for rapid bigger changes of set of active genes than just a few small mutations in the active genes can do?
That why they are called junk DNA and not trash DNA, because at least part of it is ready to be reused later.
Some parts will be to control which genes are actually activated, some might even only be necessary to determine how the DNA is folded to determine what genes get more exposure, but the thousends of broken and useless copies of genes around in the junk DNA surely also have the function of collecting mutations until they might get by chance to something useful one day.
While even proteins coded by genes many parts are just filler, the important parts do not allow much changes to still get a surviving organism. You do not want too much mutations on the active genome, or you waste too much with sterile mutants. But to get something truly different, you often need to do many changes at once, and the chances to get there with only active genome are practically not there.