Mainframe security is simple to understand and extremely intuitive to apply. Some might find it hard to believe that the mainframe security model can be easily integrated with other platforms and applications.

As everyone in the IT industry knows, a system’s security level depends strictly on the possibility of attacks from unknown and unplanned access. This kind of access is sometimes the singular cause of viruses and malware. These unexpected attacks have companies scrambling to find innovative solutions to help them avoid security breaches and improve the security postures of their IT environments.

For all these reasons, many professionals think protecting IT systems is very complicated and difficult to achieve. Why do they have this perception? In most cases, distributed systems have very complicated access control rules, requiring several actions to ensure full and secure protection.

Security-Rich by Design

An IBM white paper written a few years ago illustrated just how easy mainframe security is. The described mainframe was “security-rich by design,” which means that the hardware of the product, the IBM z Systems mainframe, is equipped with all the basic metadata required to enable the use of any software interface for protection definition.

The basic metadata is very intuitive. We can summarize its definition by asking these three questions:

• What has to be accessed? These resources can be files, transactions, programs, databases, etc.
• Who needs to access it? These are known as users.
• How strong should that access be? These are known as access types, such as read, update, alter, etc.

The hardware of mainframe servers provides this metadata as a model for every software product to use without the need for specific definitions.

This unique hardware design feature fulfills the “security-rich by design” characteristic. Any specific resource, user or access type can be defined under the corresponding model by using a specific access control tool. In fact, the hardware includes protection rules for each of the three types of metadata, so any protection definition always follows the same native rules.

User Identity and Authentication

Amazingly, this simple model saves the mainframe from attacks. A user can come to IBM z Systems through direct mainframe access, a web interface, an external program or an emulator, among many other ways. No matter where the access originated, however, the user is obliged to follow the mainframe rules and must be identified by the hardware itself to perform any final action on the system. The user must make a final request to access a resource in a specific access type. The system will reject this request if the user does not have the correct authority.

Let’s look at a quick example. Suppose the web user [email protected] wants to obtain a list of objects included in the specific IBM DB2 for z/OS database DBCUST.

After being authenticated in the web application, the user transaction follows a specific path, passing different distributed objects, to obtain data from the mainframe. Because the system is security-rich by design, the web user is almost always unable to successfully close the end-to-end transaction without an associated z System user ID (an authorized user profile).

The parameters of each specific user ID are built-in requirements specified during the development of mainframe parts. To ensure the mainframe can recognize the user ID and apply the correct authorities when an application is executed, applications should be developed considering the right associations. Every other mainframe access mechanism besides the user ID, such as pass tickets or digital certificates, must follow the same rules specified during development.

So, if we suppose the associated z System user is NICO, and NICO is authorized to access the resource DBCUST with a READ access type, then final data can be sent back to the user.

Learn More About Mainframe Security

Even as mainframe technology continues to innovate and extend access controls to modern authentication methods, such as ID cards, bank cards or fingerprints, these will always behave as unique identities for any user.

To learn more about mainframe security, check out “Security on the IBM Mainframe,” an IBM Redbooks publication.