Much as I try to understand, I think Sun's freeing up of Solaris is a bad bad move. Solaris was their cash cow!

If I were in the hot seat I would have kept Solaris up and running as paid-for-software with new innovations (like dtrace) and simultaneously have gone for:
1. new product offerings in grid computing domain and
2. ramped up java based service offerings

Also the low cost of support for Solaris is a wrong strategy.

Unless it can make revolutionary innovations in the chip design of some new high performing distributed system (plug and play), I think we will continue to see tepid growth(?) of revenue in future.

The problem with open sourcing is that not every idea is suited for open sourcing. It is a strategy not a one-size-fits-all solution.
I wouldn't open source my cash cow, even if it had a rapidly shrinking market. It would simply look elsewhere for ramping up the revenue stream.

Sun failed to cash-in with Java. (It also failed to cash in with promising technologies like Jini and Jxta.)

BTW: Sun doesn't get the IDE market. Traditionally their offerings in the Java space has been sub-standard. I am not sure why it keeps on trying again and again.