Nice to see that this thread really maintained its original intention. That rarely happens.
Only a slight advantage given by the engine?!

It's quite a logical defense. White needs c6 to get his Queen into the defense of his King, while the variation also gets tempos with the threats Nxd6 and Qd8+. Yet you see, the attack comes over the dark squares, and it's pretty difficult to defend against... as the Rooks are awful in the defense.
While black prepared his attack, if you moved your queen to b6 and planned to meet d4 with Nc5, then I think you're a chess genius. Black also has the option of 35...Nxf2 which looks promising. I don't think a human would play this way in the moves leading up to this position. After the king is close to mate, sure, players could calculate these things because almost all other moves lose.
It is important to add that this pawn structure is the Maroczy Bind in reverse. In other words, Black is exercising the Maroczy on White. An additional refinement is that it is the c-file that is half open for the defensive side instead of the e-file. It is much more difficult to break the bind with only the c-file being half open.
In order to break the bind White needs to use the squares d4 and f4 especially with the White N. Black will seek to deny White access to those squares.
Black's strategy of keeping the position closed especially in the center is consistent with not only masking the disadvantage of being down the exchange, but also the closed center facilitates a flank attack on the kingside.
In my opinion Black should win this game or White might maneuver a difficult perpetual check draw.
Most of the other analysis posts were excellent.

It's a maroczy structure (reversed) If you shift all of black's pawns one file to the left, and add a white pawn on d3.
You are technically correct. However, the missing d3 pawn is because White is down a pawn, and it makes the defense of the e4 square that much more difficult, sometimes tempting White to play the weakening f3. Additionally, the fact that the pawn structure is shifted one file over towards the kingside where the White King is castled makes it even more dangerous.
In other words, what is important is recognizing the pawn structure and the advantages and disadvantages of modifications to the pawn structure.

Why worry what the engine thinks.
Do yourself a favour and uninstall it, that way you might learn something
It is important to add that this pawn structure is the Maroczy Bind in reverse. In other words, Black is exercising the Maroczy on White. An additional refinement is that it is the c-file that is half open for the defensive side instead of the e-file. It is much more difficult to break the bind with only the c-file being half open.
In order to break the bind White needs to use the squares d4 and f4 especially with the White N. Black will seek to deny White access to those squares.
Black's strategy of keeping the position closed especially in the center is consistent with not only masking the disadvantage of being down the exchange, but also the closed center facilitates a flank attack on the kingside.
In my opinion Black should win this game or White might maneuver a difficult perpetual check draw.
Most of the other analysis posts were excellent.
Actual Black's strategy is to break through the dark squares on the King's side supported by a Knight on –most likely– e4, as White lacks the natural opponent against Black's Bishop, and White Rooks lack the space to work effectively against Black's minor pieces.
This sort of positions usually come from the QGD, but there White trades two minor pieces, making Black threats more difficult to materialize, and Black's f pawn is on f7.
There's also this vague conception of material advantage. Pieces have different values based on their expected mobility and area of attack. Should one file be opened in the Queen's side, the evaluation would be quite different, as White Rook's would have roads to get into Black's position. But, in the original diagram there's no foreseeable way to get an open file in time, thus White is effectively playing a piece down. All because of the pawn structure. And you can see it in the last diagram, one of White's Rooks is doing nothing.
Precisely on point. Booyaa!!

"Then, Black's position is better, as it can be improved while White's can't."
Hmm, well regardless of who is better, I can't really agree that white's position can't be improved... in fact, kind of my point as of late was that it seemed like black was the one waiting for white to do something, despite having a visually pleasing position. I mean, as said, white can simply get his pieces ready for some kind of pawn break -- white can play moves like Qd3, Re1, f3, and e4, moves which clearly improve his position. Maybe black has a concrete response to that, but to be honest, that seems to have nothing to do with white not having improving moves, and more to do with black having an effective counter to white's plans, if he does.
I'm not saying your analysis overall isn't interesting to read or isn't insightful, it is, but the part about a "lack of improving moves" doesn't seem to apply here, especially if we're talking about white.
That kingside plan seems interesting. It does seem smarter to just try to soften up white's pawn structure first, and then maybe get a knight to e4 and think about sacrifices on the white king some day, rather than just being a barbarian and trying to play for ...g5 and ...f4, which really opens up black's king. But, I don't think white has to just wait for black to attack him.

"If you play White and try for a "who crosses the finish line first", it's likely you'll lose as Black is more advanced into White's King side. Time is of the essence."
I don't think you generally have to, or even should, go purely one way or the other, like purely trying to stop your opponent's plans, or purely trying to further your own plans. There are plenty of openings, like the english, or the king's indian, where white does a combination of attack and defense (and black for that matter). In the english, when black does a king's indian setup, white will push on the queenside a bit, but at certain moments he will also be looking for ways to stifle black's kingside activity, and/or find ways to simplify, when his queenside ideas will gradually become more pertinent.
And so, of course, you can't just say, well black's goal is the kingside in that line, which is more valuable than the queenside, therefore, black's position/plan is better. It's more complicated than that because while black's plan results in greater potential gains, it's also more difficult to achieve, and is often much more susceptible to simplifcation.

"as the Rooks are awful in the defense."
Have to agree, there. Clumsy and poor defenders.
Something I'm a bit curious about, too, is if white can sometimes argue that he benefits from black doing an ...h5-h4 push, if he tries to get his king to g2 and (after ...hxg3 hxg3) just get the open h file, satisfying one of the problems white supposedly had of no open file.

"In other words, what is important is recognizing the pawn structure and the advantages and disadvantages of modifications to the pawn structure."
Yeah. Haha, this makes me think of an interesting possible exercise. Do you think it could be a useful exercise to take some position and then ask yourself "if black's pawn was on c5 instead of c6, would this help black's position or white's position?" Stuff like that? And just ask that same question for all of black's pawns. If his a pawn is on a6, ask yourself, who would it benefit if the pawn was instead on a7, or on a5. If you find new opportunities with those kinds of changes to the position, that can show you how your current pawn structure is compromising your opportunities. I don't know, it seems like an interesting exercise.
There are about 15 characteristic pawn structures that result from openings. There are well known middlegame plans of attack and defense for each of those characteristic pawn structures. Transpositions and knowing when they have occurred on the board are good guides to recognizing what characteristic pawn structure is on the board.
Recognizing and Knowing how to play a characteristic pawn structure from both sides is more important than recognizing and knowing what opening name usually results in what characteristic pawn structure. In other words, knowing the terrain (hills, valleys, and mountains) of the pawn structure is more important than knowing the geological history and origin of the terrain.
No matter what variant of an opening you have in front of you as a position. It is more important to recognize the pawn structure or its cousin than to know the opening it originated from. It all comes down to one word, "perspective".

There are about 15 characteristic pawn structures that result from openings. There are well known middlegame plans of attack and defense for each of those characteristic pawn structures. Transpositions and knowing when they have occurred on the board are good guides to recognizing what characteristic pawn structure is on the board.
Recognizing and Knowing how to play a characteristic pawn structure from both sides is more important than recognizing and knowing what opening name usually results in what characteristic pawn structure. In other words, knowing the terrain (hills, valleys, and mountains) of the pawn structure is more important than knowing the geological history and origin of the terrain.
No matter what variant of an opening you have in front of you as a position. It is more important to recognize the pawn structure or its cousin than to know the opening it originated from. It all comes down to one word, "perspective".
Quoted from the introduction in Soltis' book right? Or it's at least similar sounding.
Anyway, it's important to see that the board is not symmetrical due to the kings on e file and queens on d file. I think Soltis mentions that in his book too. For example the king's gambit and queen's gambit are completely different in nature.
The OPs position is NOT a transposition into a Maroczy structure. The plans are different because e file vs d file. As far as attacking on the kingside I'd say it reminds me of a stonewall structure.
Colors reversed Maroczy (with white's d pawn missing) would look something like this:
There are about 15 characteristic pawn structures that result from openings. There are well known middlegame plans of attack and defense for each of those characteristic pawn structures. Transpositions and knowing when they have occurred on the board are good guides to recognizing what characteristic pawn structure is on the board.
Recognizing and Knowing how to play a characteristic pawn structure from both sides is more important than recognizing and knowing what opening name usually results in what characteristic pawn structure. In other words, knowing the terrain (hills, valleys, and mountains) of the pawn structure is more important than knowing the geological history and origin of the terrain.
No matter what variant of an opening you have in front of you as a position. It is more important to recognize the pawn structure or its cousin than to know the opening it originated from. It all comes down to one word, "perspective".
Quoted from the introduction in Soltis' book right? Or it's at least similar sounding.
Anyway, it's important to see that the board is not symmetrical due to the kings on e file and queens on d file. I think Soltis mentions that in his book too. For example the king's gambit and queen's gambit are completely different in nature.
The OPs position is NOT a transposition into a Maroczy structure. The plans are different because e file vs d file. As far as attacking on the kingside I'd say it reminds me of a stonewall structure.
Colors reversed Maroczy (with white's d pawn missing) would look something like this:
It is all about perspective. If it helps you to see it as a stonewall structure, then that is your perspective. Practice with perspective about a position is at the foundation of selecting a good move.
An seemingly unrelated example comes from the endgame K+2B vs. K+N. For 300 years, even after extensive analysis before computers, this endgame was believed to be a draw. The computers with mind numbing number crunching gave us a new perspective. With very few exceptions this endgame is a forced win for the K+2B.
Our brains gave us a new perspective. It turns out that every time that the Bs traverse to a new diagonal from the edge of the board they gain a tempo. The technical details of executing a win requires some 200 moves. This of course is well beyond the 50 move rule. The purpose of this example is that flexibility in our perspective is paramount.

Pawn structures shouldn't be thought of as symmetrical across the vertical axis because the king and queen.
For example this
Is not the same type of game is this

"This is a matter of concept: You modify the pawn structure to increase your piece activity, not your opponent's. It makes a lot more sense to modify the pawn structure on the Queen's side, as White has space advantage over there, but White lacks time as Black's too close to White's King."
Yes, of course f3 is a weakening move, but it also provides benefits to white. It's not strange for an idea to have both good points and bad points to it.
I would say, fundamentally, e4 would help white much more than it would help black. It takes away part of what was so great about black's pawns on d5 and f5, how they gained space, and locked the position for the rooks. Now there are plenty of openings in the position, (if you imagine a lot of exchanges happening on e4 for example) and it starts to look like, you'd just much rather have a rook. Also, white's bishop would have much more influence than it does currently, locked in by black's d5 pawn. Or perhaps it would exchange itself for a knight, which, although it weakens the king, also trades a passive piece for an active one, and simplifies, so it's worth considering.
The problem with the plan is that it weakens white's king position. White benefits in the center, but he has to weaken his kingside, and maybe black can take advantage of that. It is nice that white is preventing ...Ne4 with the move f3, but g3 is definitely a weak spot. But this is chess, as I said. Whether f3 and e4 works may depend on black's setup, or other concrete factors in the position. I'm just saying, it's something that white would like to have.
""I don't think you generally have to, or even should, go purely one way or the other, like purely trying to stop your opponent's plans, or purely trying to further your own plans. There are plenty of openings, like the english, or the king's indian, where white does a combination of attack and defense..."
None of those concepts apply to the OP's diagram. If someone with 10 dollars in his pocket walks into a restaurant with people having 1000 dollars in each of theirs, he can only order up to 10 dollars, regardless of the averages."
Of course they apply. White can decide what is the perfect amount of time to spend gaining space on the queenside, and how much time should he spend concretely reacting to black's threats.

It's like in the QGD exchange. It can sometimes be hard to decide whether to go for f3 and e4, or b4-b5, since they both look tempting and sensical :) I would say with the extra exchange, you would like to play f3 and e4, but whether you can deal with the kingside weaknesses that come with it is a question white needs to answer. But remember, I was addressing your comment that suggested white has no plan or way to improve his position. That's not true. He has plenty of ways to improve it. The question is, is black's kingside initiative more important than white's trumps.
I think OTB black's attack is very dangerous, but look at this completely ridiculous computer defense putting the queen and knight far away from the king. It looks like black should win, but engine says 0.00