Fwd: framework question

[Simon Coulter <shc@xxxxxxxxxxxxxxxxx>]

From the right ID this time ....

Begin forwarded message:

> On 07/07/2004, at 10:29 PM, Joe Pluta wrote:
>
> > This is the type of bollix I'd really like to keep off this list.
>
> As opposed to the bollocks you spout under the guise of <MODERATOR 
> ALERT>?
>
> > Having journaling and commitment control may indeed be more reliable
> > than not having them.  However, there is a price to pay in terms of 
> > disk
> > space and performance.  That price is essentially the cost you pay for
> > insurance against hard disk crashes.
>
> I really should know better ... however ... you don't get something 
> for nothing. Even doing it the roll-your-own way has a cost. I don't 
> know about you but DASD is much cheaper than my time--even at IBM 
> prices.
>
> DASD space usage can be managed by changing journal receivers 
> regularly. DASD arm usage can be managed by putting receivers on 
> dedicated drives. The 5% or so CPU overhead I would accept as a 
> reasonable payment for guaranteed database integrity.
>
> But I forget. You could write a secure safe reliable transaction 
> processor in less time than the rest of us could eat lunch so the 
> programmer cost would be negligible. (Oscar was right about sarcasm.)
>
> > So, once again, this is NOT a philosophical issue, it's a business
> > decision.  Does it make sense to spend the money on commitment 
> > control?
> > How often does your machine have a hard crash?
>
> Exactly, it is a business decision. How much does it cost your 
> business to allow programmers the luxury of reinventing integrity 
> mechanisms when the system already provides such features. How much 
> does it cost your business when a crash occurs? How long can you 
> afford to be down while someone verifies the database is intact?
>
> This is another example of "past behaviour being no indication of 
> future performance". Just because a hard crash is unlikely doesn't 
> mean you shouldn't plan for such an event nor do what you can to 
> minimise the impact. Especially when a small amount of development 
> investment in applying commitment control and a small amount of cost 
> for DASD and CPU can guarantee recovery to a known point.
>
> Regards,
> Simon Coulter.
> --------------------------------------------------------------------
>    FlyByNight Software         AS/400 Technical Specialists
>
>    http://www.flybynight.com.au/
>    Phone: +61 3 9419 0175   Mobile: +61 0411 091 400        /"\
>    Fax:   +61 3 9419 0175                                   \ /
>                                                              X
>                  ASCII Ribbon campaign against HTML E-Mail  / \
> --------------------------------------------------------------------