Hello,
Initial "disclaimer", I work in IBM. I'll try however to be balanced, especially since I'm more interested in clarifying a few points than in engaging in some sort of competitive bashing.
For the record though I'll say that I like Solaris and business imperatives apart Sun is/was a company that interested me.
IBM's offerings with their overpriced hardward,
Really depends on how you do the math. Individual systems can be more expensive, but then again they generally do a lot more in terms of processing power. Of course, "processing power" can be again measured in multiple ways, which is why you'll find a lot of contradicting information. One thing to bear in mind though is that, for example, the IBM Power Blades are quite competitive, being similarly priced as the ix86 ones. The higher you go in terms of vertical capacity of growth, the pricier it is, but that's the same in all vendors.
ancient lineage
I'm not sure what you're intending to say here, most Unix vendors have an ancient lineage (Solaris itself is a BSD/System V mix, a bit like AIX). If you're referring to a supposed lack of innovation, well, POWER6 still has the edge in terms of processing power and POWER7 is just around the corner (IBM won the DARPA bid against Sun btw). AIX 6 introduces a lot of new stuff which you are probably not aware. I'm not sure how is the Sun situation in terms of chip manufacturing. I know about the highly threaded CPUs, etc, I am just commenting on the possible perception that looms in the air with the Oracle acquisition.
how about a free x86 version, IBM? no, then fuck off!
While I understand that it would be interesting in general terms, it doesn't matter in terms of judging the fitness of the OS for the market we are talking about.
After that, I have a hard time figuring out why anyone would favor IBM's LPARs over the much more efficient, and easier to manager Solaris 10 Zone offering.
They are quite different concepts though.... a LPAR is for most purposes a separate server, with a level of isolation that exceeds Solaris zones. They don't even compete in the same area. A critical problem in the Solaris kernel that is supports 10 Containers will mean death to all of them (correct me if I'm wrong). You can do whatever you want to an LPAR that it won't affect any other LPAR. This with the added benefict of dedicated OR shared hardware, dynamic CPU and RAM entitlement via policies, etc, etc. It behaves a bit like z/VM.
The only comparison with Solaris zones are WPARs, Workload Partitions, introduced with AIX 6. They share a global kernel and behave in a similar way to Solaris Containers, give or take. They are lighter in terms of creation, etc, but with less isolation. I'm sure that there are arguments pro and against each of them, but in terms of use they can be compared. Not so with LPARs.
Don't get me started on HP
HP-UX is a solid UNIX OS. Of course, it isn't as "sexy" as Solaris (like AIX also isn't I guess), but again what matters for most is if it's stable and manageable. HP also has different virtualisation offerings (nPars, which work at the physical level a bit like Sun Domains IIRC, vPars which are lighter weight and share the same hardware, IVM which is sort of like VMware in Itaniu, etc). I always was an admirer of the Alpha architecture, and respected PA-RISC. In personal terms I don't especially like Itanium, *but* this is a personal thing.
Ever heard of ZFS or DTrace?
Quite interesting features. I especially like DTrace.
I will disclose that I am a three-time ex-Sun employee/contractor who has also seen inside the belly of IBM. Solaris will bury AIX. And you can take *that* to the SAN and store it!
Well, we're both a product of our surroundings I guess :) I disagree that Solaris will "bury" AIX of course. You might enjoy Solaris more - that's quite reasonable - but in the end this things are more about the business sense they make than anything else (and I'm not saying that Solaris doesn't make business sense).