22.12.2012 Views

Front cover - IBM Redbooks

Front cover - IBM Redbooks

Front cover - IBM Redbooks

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

138 Lotus Security Handbook<br />

Note: It is theoretically possible on some platforms to provide some level of<br />

service separation in a single host by running the services under different<br />

non-root-privilege accounts. We say separation within a single host is<br />

“theoretically” possible because in reality it is complex to configure properly,<br />

meaning it is prone to errors that can lead to unintentional vulnerabilities. For<br />

example, on UNIX systems it is possible to isolate daemons (background<br />

processes or services) such as bind using chroot; however, if not configured<br />

properly, an attacker may still be able to break out of the “chroot jail” and affect<br />

other parts of the system. A simple Google search on “chroot breaking out”<br />

locates numerous Web articles, such as the following reference:<br />

http://www.bpfh.net/simes/computing/chroot-break.html<br />

We do not recommend using service separation on a single host as a network<br />

security method because service isolation cannot be guaranteed.<br />

The focus of our multi-zoned architecture will be on the placement of servers,<br />

applications, and services to provide separation. As previously mentioned,<br />

determining which zone a server or service belongs in can be based on<br />

sensitivity of data and risk.<br />

The ideal architecture can provide granularity of access to data as dictated by<br />

the requirements of the business application owners. The requirements must be<br />

consistent with corporate security policies. In order to provide granularity,<br />

flexibility, and future extensibility, we need to determine a reasonable number of<br />

security zones we can classify and locate our host systems into. The four<br />

categories, or types of security zones we define in our model are:<br />

1. Internet zone<br />

2. Proxy zone<br />

3. Data access zone<br />

4. Intranet zone<br />

As we define these four types of security zones, note that many of their<br />

characteristics are derived from the connection restrictions we must have for a<br />

given zone type to or from another type of zone. So as we define the zone, we<br />

list some recommended guidelines that we use later for more specific policies for<br />

connecting to the zone. Also note that we use language in our policy guidelines<br />

such as “will have”, not words like “should have”. As you formulate your own<br />

policies, be aware that they must make a clear distinction between what is<br />

permitted and what is restricted, and do not allow critical restrictions to appear as<br />

if they are something optional. In other words, avoid any ambiguous language.

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!