The biggest mistake I think people do when trying to understand time dilations is to assume that there must be some common 'standard' to it . There is no such thing, the only 'standard' your universe gives you is your own local definition of 'c' and your local 'arrow of time' measuring out your unvarying lifespan. You can extrapolate that standard to include everyone in the universe of course, seeing that we all find 'time' unvarying from our local frame of reference, as well as 'c' always will give us the same definition, if measured directly and locally.

And if you do I agree

We live as entities in a sea of unvarying radiation. most of us moving 'uniformly' in space. That radiation is a 'clock' of a sorts, it always ticks the same locally. But motion distort the 'room time' it 'ticks' in. And the only way we have defining motion is 'expending energy', aka accelerations. Without knowing who accelerated relative whom, defining a 'time dilation' becomes meaningless.

Any possible time dilation you ever will be able to define must have a common origin, a same 'frame of reference' from where you can set the 'clocks' to being 'approximately' the same. Without a common origin, there is no way to prove who becomes 'time dilated' relative whom.

Imagine a situation where we have two frames both moving uniformly with you looking at both from Earth. Deciding to define Earth as your 'inertial frame', at 'zero' speed, you now are able to give both of those object a unique 'speed' relative your 'newfangled definition' of 'zero'. So you state that 'A' is moving with double the 'speed' of 'B', according to your observation relative earths 'zero', and so must be aging 'slower' than 'B'. And you are right, relative your definition of Earths speed being 'zero'.

Now be 'B' looking at 'A' . From the point of 'B' moving uniformly you are free to define 'B' as being 'still' too. You then define 'A' as the only one moving, therefore having an even slower clock than the one you found before, using Earth as your 'pivot of change' (only as an example, okay:)

Now be 'A', looking at 'B' moving uniformly. You make the same assumption again and find that it must be 'B' that is the one moving, therefore now having the slowest clock. And all of those definitions are correct.

Why are they correct? Because you do not have a same origin. Without a origin and a clock to set all further 'time dilations' against, any definition you make will be arbitrarily. And if measured looking at the others clock, you will find all clocks to go slower than 'yours', no matter from where you chose to stand looking.

It doesn't mean that the 'arrow of time' is a illusion though. As I see it, it's because of it not being an illusion, instead having a 'clock' of a unvarying 'c' locally that those effects can exist. Think of the radiation between 'frames of reference' as something 'invariantly ticking', the 'room time geometry' contracting or magnifying in your relative motion, versus all other 'frames of reference'.

Also remember that any 'uniform motion' only can exist as a definition (comparison) between 'frames of reference' in Einsteins universe. And he's right there. You need at least two objects, then a space (distance), and a 'clock' ('c') measuring even durations, to find a motion. Your universe gives you one clock 'c', but, with it a indefinite amount of objects, all of them in 'different' uniform motions, as defined from your point of view.

Looking at it from that perspective you will find everything 'time dilated' but, only measurable, when you can define some common origin from where you set your clocks. And there is no way around this, you need a same origin, and you also need to meet up with the twin, to define that 'time dilation'. And, Nota Bene, you now also will have to introduce a acceleration to get that 'time dilation', and with that acceleration you also will introduce 'gravity' in the accelerating frame, and so a locally observable change defining whom the time dilation 'belongs to'.

Although 'time dilations' is a reality, your invariant clock as I define 'c' as, also will be a reality locally. And that fact will give you a same life measure inside your 'frame of reference' no matter where you are, or how fast you go. And that one is extrapolate-able to all other frames too, conceptually.

==

Room time geometry is just the way the room (time) will distort relative the observer. You look at someone traveling finding him 'slower', that I call the travelers 'room time geometry', as defined from you. Defined from him it will be what he observes outside his own 'frame of reference' that is distorting instead. And that's also why I define it as being a 'relation', just as a uniform motion must be.