Re: [Bug-gnupedia] Architecture Questions

[Bob Dodd]

--- Rob Scott <address@hidden> wrote:
> I think it is a better idea to have the actual storage
> centralised into one box, and the page serving and
> interpretin should indeed be split up into many
> identical boxes, which store no data themselves.   It
> is very hard to implement a system which distributes
> the data among various machines, as it has to be
> copied to the appropriate machine after being
> submitted, and then to all other 'mirroring' machines,
> which all adds an overhead of processor time, network
> activity etc...
> 
> I think its far easier to use just one actual db
> machine and to ensure it is kept up 

With a decent database platform, it's not *your* problem. I would
expect mirroring of the database (I'm talking about within one physical
location, not "mirrors" of the encyclopedia around the world) to be
handled internally by the database itself; it's primarily a set-up
problem. Exactly which mechanism is used for the mirroring is chosen by
the people who wrote (say) mySQL. However they do it, as a db
administrator, or as a piece of software querying the db, you will
still see one logical view of all the data. Of course the fun comes
when your db is too mickey mouse to do mirroring transparently... 

That's why I was asking what the mySQL apporach is.

/Bob Dodd









__________________________________________________
Do You Yahoo!?
Yahoo! Auctions - Buy the things you want at great prices. 
http://auctions.yahoo.com/