IT Passport Examination (IP) | Technology Questions on Virtualization, Backup, Reliability, and Authentication 10

A delayed discovery can mean that the latest backup already contains the corruption. Retaining multiple generations increases the chance of returning to a healthy point before five days ago. Repeatedly overwriting a single latest copy can destroy that recovery point. Retention periods and restoration testing should also be designed.

Live migration moves running virtual machines to another physical host and can minimize interruption caused by maintenance on the original host. The platform's requirements for processor compatibility, networking, storage, and other factors must be verified. Power removal or data deletion would damage availability and data protection.

Container virtualization isolates applications and their dependencies while sharing the host OS kernel. It is generally lighter and faster to start than virtual machines that boot a complete guest OS for every environment. It has constraints when workloads require freely mixed operating systems with different kernels.

A duplex system uses a primary system for normal processing and switches to a standby system after failure. It differs from a dual system, where two systems perform the same processing and compare their results. Switching time and potential transaction loss depend on the standby state and synchronization method.

Hot standby keeps the secondary system close to an operating state and reflects primary-system state to support rapid failover. Cold standby generally requires startup and preparation after failure, making failover slower but often reducing standby cost. The choice should balance the required recovery time and cost.

The mechanism refuses an unsafe operation—heating with the door open—despite user input, so it is foolproof design. Fail-safe design moves equipment to a safer state after a failure. Its focus differs from preventing an invalid user operation from taking effect in the first place.

The redundant power supply lets the server continue its required function after one supply fails, making this fault-tolerant design. It does not attempt to guarantee that failures never occur; it maintains service when an anticipated failure occurs. The tolerated number of failures and common-cause failures must also be considered.

A stricter match threshold tends to reduce FAR, in which an impostor is accepted, but it can increase FRR by rejecting more genuine users. The threshold therefore requires a balance between security and usability, including retry and fallback procedures. Both error rates do not necessarily reach zero together.

A timestamp from a trusted time-stamping service can demonstrate that the target data existed at that time and has not changed since. A local filename or modification time can be altered by the user and is not third-party proof of time. A timestamp does not establish that the document's substantive content is correct.

A certification authority can revoke a certificate before expiration after events such as private-key compromise. A CRL provides information about revoked certificates, so a revoked certificate should not be trusted for server authentication. Some implementations use other revocation mechanisms such as OCSP.