1–10 of 31 posts
Thanks SunSolaris? Linux? FreeBSD?
My 2 cents: Oracle is going to drive Solaris into the ground (and OpenSolaris doubly so). It's days were already numbered, like all proprietary Unixes, but while before I was betting on Solaris to be the last one standind, now I think it will probably be AIX.
On what do you base this? Larry is spending more on SPARC and Solaris than McNealy ever did! It's possible that Solaris will become less suited to (for example) HPC over time but for anything related to commercial data processing, including Java and web serving, expect for it to continue to occupy a leading position among the Unixes.
If you're unfamiliar with both, Linux systems are easier to get used to, in my experience.
Ask yourself where your bottleneck will be, then ask yourself if the OS is going to make any difference there. The answer is very likely no.
No modern OS is going to give you any significant performance/cost advantage over another in serving regular web apps. Most of your trouble will be in setting it up and maintaining it, so go ahead and pick whatever you're most comfortable, or, if you're not comfortable, ask someone who are (and that you trust).
Also, don't pick CentOS, unless you need RH compatibility.
None of your requirements, at least when they're as unqualified as they are, can help you to pick an OS. Ask yourself where your bottleneck will be, then ask yourself if the OS is going to make any difference there. The answer is very likely no. No modern OS is going to give you any significant performance/cost advantage over another in serving regular web apps. Most of your trouble will be in setting it up and maint…
I saw that oracle are really emphasizing how their OS has much higher performance than linux I wondered if its correct or if their benchmarks are good: http://www.oracle.com/us/solutions/performance-scalability/b... http://jp.sun.com/solutions/infra/eda/pdf/solaris-linux.pdf
So I wondered if i'm missing some tens of percents in performance by not choosing Sun Solaris...
Whatever you choose, stay clear of Solaris. Especially OpenSolaris (that's a dead-end). If you're unfamiliar with both, Linux systems are easier to get used to, in my experience.
My 2 cents: Oracle is going to drive Solaris into the ground (and OpenSolaris doubly so). It's days were already numbered, like all proprietary Unixes, but while before I was betting on Solaris to be the last one standind, now I think it will probably be AIX.
Oracle is going to drive Solaris into the ground On what do you base this? Larry is spending more on SPARC and Solaris than McNealy ever did! It's possible that Solaris will become less suited to (for example) HPC over time but for anything related to commercial data processing, including Java and web serving, expect for it to continue to occupy a leading position among the Unixes.
I don't see it "running into the ground" - people will be using it for many years in the future. However, the supply of people picking it for new projects probably just dried up quite a bit. People will 'grow up' knowing how to work with Linux, because that's what they tried at home, at school, what they learned on the job, and what their friends know. Those network effects mean that it'll mostly be sensible to keep switching to Linux over time, gradually, unless Solaris suddenly sprouts something that makes it a must-have, which I don't see.
My 2 cents: Oracle is going to drive Solaris into the ground (and OpenSolaris doubly so). It's days were already numbered, like all proprietary Unixes, but while before I was betting on Solaris to be the last one standind, now I think it will probably be AIX.
Oracle is going to drive Solaris into the ground On what do you base this? Larry is spending more on SPARC and Solaris than McNealy ever did! It's possible that Solaris will become less suited to (for example) HPC over time but for anything related to commercial data processing, including Java and web serving, expect for it to continue to occupy a leading position among the Unixes.
Oracle's strategy with Solaris is going to be to squeeze every last penny from it, and when it dies it dies. Killing OpenSolaris and making it clear that open source releases of the code base will be done only "at some point" after the commercial release are just the beginning of that.
They will push Solaris in the same places where they currently push Oracle DB. Those places tend to be of very little interest to 99% of the HN croud. Java and web serving are not the kinds of things that justify stratospherically overpriced software licenses...
As for SPARC, I seriously doubt they will pour significant money into it. Trying to compete with x86 is akin to setting $100 bills on fire at this point. They will no doubt keep the Sun brand alive and use it to sell x86 based machines (which is the strategy Sun itself started of course).