Re: backup disk format

Peter Nunn <pnunn_at_pncomputing.net> wrote

> I'm after the backup format because I'm thinking about producing a
> product that can open backup files from all the major database
> manufacturers, view their structure and report on that (sort of an
> integrity scanner if you like).
>
> I've come to the conclusion that the native formats are going to be
> hard, but am wondering if the backup formats would be worth a look.

Oracle does not have a separate backup format. Nor does SQL-Server or Informix as far as I can recall.

In all these products, you have binary data files making up a database. In some cases, raw devices.

Backup means making a _physical_ copy of that file or raw device to backup media (usually also compressing it).

It also gets a tad more complex. In Oracle we do not just backup a copy of the physical database data file. Instead we also backup binary files called Archive Logs - these contain the transactions that have been applied. And can be used to reconstruct the database files using an older physical backup of those files.

Then there are logical backups such as exports and sql-dumps. But these are very seldom use as the prefered method to backup an entire database.

> I'm going to try and do a binary dump of a backup, but if someone has
> already cracked it, it would save a lot of time.

Besides the legal issues with reverse engineering these formats, there's also the issue of these formats changing between product versions.. and becoming increasingly complex. Cracking the data file format of Oracle 9 will take many man months.. by which time Oracle 10G will be released..

What business benefit will you be able to provide with such a product? Even if you can sell it commercially without legal repercussions? How will you convince customers that your product can correctly process/check the db vendor's physical data files when said vendor is not a partner and did not provide you with the actual data file formats?

Customers want to have data security. They do not want to "view" the contents of the data they secure - that is the job of the db engine. And it does a fine job of that by providing transactions and data integrity.

Backup software has the ability to provide backup integrity (after writing a block to backup media, reading the block from that media and comparing it with the block in memory to ensure integrity).

--
Billy