Interesting situation for both sides

The following position occurred after 17 moves. I played with the white pieces against a not-so-well played Sicilian Dragon. After my 17th move, my opponent had about 1 minute and 30 seconds left on his clock. His time eventually ran out, as he was unable to come up with a plan. What would your plan be in this position as black? As white?

Just looking at it for 5 seconds my plan as white would be maybe Nc3, Nc2 and then torture black some because he has no pawn breaks and must suffer.  What is this, the bind right?

Black has to get in b5, f5 or d5 at some point if he doesn't want to lose.  And by must I mean it doesn't matter if it's a sacrifice or not because right now he's dead in the water (such is the bind).  a5 was obviously a very poor choice.

Oh it's black to move, then I'd look at Bxd5... I guess that loses the a pawn, black may not be that desperate  yet... maybe Nh5 or Ne8 with f5 immediately.  That also may be rushed... again this is just at a glance... to be so rushed probably isn't good, how about Qc8 with Nd7 which seems good.  By the way Qc8 threatens to win the bishop with Nd7...  so white would probably play...

First, Black has to derail White's threat of Nc7. If Black takes on d5, the a-pawn falls, as orangehonda pointed out. If Black plays 17...Rfc8, then an eventual ...f5 break won't be nearly as effective. 17...a4 18. Nc7 Rac8 19. bxa4 is just hellish for Black. So this leaves 17...Ne8 followed by ...f5, when Black's pieces get some life.

*waits for a correction from somebody who may actually have a clue*

Very interesting question... and the more I look at it the more I like it! In correspondence chess I can see spending quite a long time on that position. In fact, if it were blitz, I just timed out considering my options... Black has a bunch of tantalizingly interesting plans (based on opening up the a1-h8 diagonal, and putting a N on the weak d4 square) ... but none of them quite work when you calculate the details... I'm gonna come back to this one.

For a computer's perspective, my Rybka actually likes (initially, no 30 min analysis here) the awful looking a4 for black (I guess it's that bad already).  White of course plays b5 and black is just looking awful in my opinion.

Also in my main line, 19.Nd5 is just slow, 19.f4 with 20.f5 to follow is best (common attacking move, plus prevents black from ever playing f5 himself).  I saw this but giving up e5 wasn't clear to me... it's very clear to Rybka with white at +1.00 already.

But basically that's what this is -- black has to engineer an effective break with b5, d5, or f5 to get any kind of play.  If white can stop black from doing this, black has nothing but to wait around to die.

I think the problem for black is that there's no concrete plan for him. White's play is so easy with so much space, but black has no good targets. I think white is better, but white is one slip away from disaster.

Just pointing out... white's Nc7 isn't a threat at all :) black just moves the rook.  If white then takes the nasty e6 bishop black will thank him very much for prompting his queen rook to a better square, taking off the not so useful e6 bishop, and removing a piece when black has less space in the first place.

You're right though that getting f5 in is a priority for black.  Ne8 doesn't seem the best way though.  Although I admit in my line the computer showed how white prevents f5 all together so it may be something like the best practical chance... vs non-GMs anyway .

It would do you some good to look into the Maroczy Bind, which is what this basically is.  http://en.wikipedia.org/wiki/Mar%C3%B3czy_Bind

Black actually has more versatility in his plans (he can play for any of the 3 common breaks, b5, d5, or f5) although it does depend a bit on how each side plays his pieces.  But in any case white does certainly have the advantage.  White isn't one slip away from disaster, he simply has to play with patience and be willing to maneuver.

Black loses the a5 pawn after he moves the queen's rook away. Are you saying that Black gets too much compensation for the pawn?

I like to always consider piece activity -- if you don't have activity then material is worth nothing... so yes actually I think black has some compensation and at the same time white will have given up two of his best minor pieces (But more importantly is left with the passive e2 bishop and dark square weaknesses that make his position resemble swiss cheese).

I guess it's a pawn sacrifice, I didn't look too far into it, I thought Rac8 (then after a pawn drops) rook back to a8 won white's a pawn, but of course white just plays Qd2... still, if I could force white to do this I would definitely take that position, black has activity, open lines, and targets on the queenside while white is left to be passive.

Take that and compare it to the position the OP gave us where black has next to nothing going for him :) -- with perfect play I have to believe white has better than to grab a rook pawn and then go into passivity + accept so many dark square holes.

Maybe I was too harsh on winning that pawn, it's not like white wouldn't be winning... or at least still have the advantage anyway.

I was watching 2 masters analyse outside the playing hall my last tourney and one of them showed a line where he could have won a knight, but then quickly reset the position saying "I didn't go for that because of course it's bad" and the other master agreed.  So I went ahead and asked what about that position made it so bad, and he said, sure I win a knight and ok my position would still be better, but it gives black counter play, why would I do that when my position here is just crushing... it's too early to cash out.

So basically I guess grabbing the pawn is fine (although of course there is some compensation), you would get to keep it -- I just think it would give black reason to hope, when in the original position he should have little to no hope .

You can never be too harsh on taking poisoned pawns.