Database Reference
In-Depth Information
Engineered Systems
Hardware and software can be combined in manageable rack-mounted and pre-installed units. This leads to more
standard installs and greater efficiency. Unfortunately it has also been leading towards DBAs having more root access,
which breaches the SoD required to secure the DB from “root” Unix. I recommend Arup's article on Engineered
Systems Administration (first link) as an illustration of the issue, along with the second link, which is from Oracle
Corp, on DB machine administration:
Big Data
We could have another whole book on big data as well. Stanford has published a great video on Cloudera's version of
Hadoop, which is available freely here:
CSA are dealing with security aspects of big data at this URL:
Big data is fast on large data sets, without spending too much money, but the data integrity has been low. RDBMS
has good integrity, so connecting RDBMS to Big Data is very interesting as it combines the speed with integrity. This
does raise security issues, as Hadoop does not really have a security model yet, and if it is connected to a HR DB—in
order to correlate employee information, for instance—then there could be a risk that an attacker could gain privileges
on Hadoop and then use that to query the sensitive data in the connected RDBMS. I suggest this URL on the Oracle
Big Data Connector as a starting point:
Additionally, the Oracle BI VM has 12c with the big data connector already pre-installed. It is 50 gigabytes but will
save time with installations. See the OBI sample application at this URL:
BTRFS
BTRFS is a Copy on Edit file system, which improves recovery from corruption as data is not deleted and updated, it
is simply copied to a new value. This improves general availability and integrity and is a significant contribution from
the Oracle Linux community. It would be interesting to spend more time looking at the security pros and cons of this
new technology. For an entertaining introduction please see this URL:
