Re: Oracle RAC for scalability or High Availability only

On 2006-03-01, DA Morgan <damorgan_at_psoug.org> wrote:
> JEDIDIAH wrote:
>
>>>No more or less so than any other method of satisfying the same
>>>overhead. One can always choose to favour the networks by just limiting
>>
>>
>> That overhead DOESN'T EXIST in a non-clustered solution.
>
> And that overhead is a known percentage of CPU and easily compensated
> for.

        I'm not talking about the memory or cpu overhed of crs. I'm talking about the fact that you need to deal with cluster technology and hardware. A NUMA machine can be treated like a fat desktop for the most part when compared to raw/ocfs/cfs/vxfs, crs and multiple nodes across a san fabric interacting with each other.

>
> It might interest you to know that the overhead of maintaining indexes
> doesn't exist in many non-relational database solutions. Planning to
> use on?
>
> Didn't think so.
>
>> A single system != a collection of systems.
>
> Correct. Single systems fail.

        RAC is not immune from this itself.

-- 
...as if the ability to run Cubase ever made or broke a platform.
                                                                  |||
	                                                         / | \

 Posted Via Usenet.com Premium Usenet Newsgroup Services
----------------------------------------------------------
    ** SPEED ** RETENTION ** COMPLETION ** ANONYMITY **
----------------------------------------------------------        
                http://www.usenet.com