Test on Computer I/O and Storage Systems
Computer I/O & Storage Systems: A Student's Guide
Test: Storage and I/O
20 questions
Question 1: Direct Memory Access (DMA) operations inherently ensure that cached copies of memory blocks remain consistent with the main memory without requiring any specific management techniques.
A. Ano
B. Ne
Explanation: If DMA writes to a memory block that is cached, the cached copy becomes stale. Similarly, if a write-back cache has a dirty block and DMA reads that memory block, it will read stale data. To ensure cache coherence, it is necessary to flush blocks from the cache if they will be used for DMA, or use non-cacheable memory locations for I/O.
Question 2: A disk with a manufacturer-quoted Mean Time To Failure (MTTF) of 1,200,000 hours (140 years) is guaranteed to function for at least that duration.
A. Ano
B. Ne
Explanation: The quoted Mean Time To Failure (MTTF) does not guarantee that a disk will work for that specific duration; it is merely the mean time to failure across many units, not a lifespan guarantee for an individual disk. The study materials explicitly state that assuming a disk will work for its MTTF is 'Wrong'.
Question 3: RAID 5 distributes parity blocks across its disks.
A. Ano
B. Ne
Explanation: RAID 5 is described as having 'parity blocks distributed across disks', which helps avoid the parity disk from becoming a bottleneck.
Question 4: I/O system design always prioritizes maximizing throughput, even for time-critical operations where latency is a primary concern.
A. Ano
B. Ne
Explanation: The study materials indicate that I/O system design can involve a trade-off between response time (latency) and throughput. Specifically, it mentions 'Satisfying latency requirements For time-critical operations' as a distinct aspect, implying that in such cases, latency is prioritized, not always throughput. Maximizing throughput involves finding the 'weakest link' and configuring the system accordingly, which is distinct from satisfying latency for time-critical operations.
Question 5: RAID systems are designed to provide fault-tolerant storage.
A. Ano
B. Ne
Explanation: RAID (Redundant Array of Inexpensive/Independent Disks) provides fault-tolerant storage by using extra disks for redundant data storage. For example, RAID 1 uses mirroring where data is replicated across N+N disks, allowing reads from the mirror disk upon a disk failure. RAID 6 offers even greater fault tolerance through two lots of parity.