| Oh, to get to the question in your post, if you're talking specifically about the systems surrounding a SAN device(or devices) on ultra large-scale deployments to ensure availability to every cluster member(hundreds, thousands, tens of thousands) - I'm not a real big expert on out-of-the-box solutions because I would never use them. First, the solution will always be extremely subjective based on all kinds of variables. Second, linux in particular has some extremely powerful constructs with which to build load-balancing and high-availability solutions. I'd be working mostly with some combination of IPVS and BGP(in the case of a country-wide cluster as a good example of where BGP could shine).
When it comes down to the actual devices themselves, I'll probably get slammed for this, but I really dig the Dell EqualLogic systems when iSCSI is feasible. | 