When configuring a storage array in a data center, which of the following RAID levels provides both data redundancy and performance improvement?

RAID 10 (a stripe of mirrors) combines the features of RAID 1 (mirroring) and RAID 0 (striping). It stripes data across multiple mirrored pairs of disks. This configuration provides excellent data redundancy, as one disk in each mirrored pair can fail without data loss. It also delivers significant performance improvements for both read and write operations, as I/O is distributed across all disks in the array. RAID 5 (striping with distributed parity) also provides both redundancy and performance enhancement. Data is striped across multiple disks, which improves read performance. Redundancy is achieved by distributing parity information across all disks, allowing the array to withstand a single disk failure.

A. RAID 0: This level, known as striping, offers performance improvement by writing data across multiple disks but provides no data redundancy. A single disk failure results in the loss of all data.
C. RAID 1: This level, known as mirroring, provides excellent data redundancy by duplicating data on two or more disks. However, it does not improve write performance over a single disk.

1. HPE Smart Array SR Gen10 Controller User Guide:

Page 30, "RAID 10 (Striping and Mirroring)": "This RAID level is a stripe of mirrors that offers the highest performance, but at the highest cost. It is a good choice for sites with high-performance requirements." This confirms RAID 10 provides both high performance and redundancy (from mirroring).

Page 30, "RAID 5 (Distributed Data Guarding)": "This RAID level stripes data and parity across all drives. This feature provides data redundancy and good performance for read-intensive applications." This confirms RAID 5 provides both features, though it specifies its performance benefit is primarily for reads.

2. Patterson, D. A., Gibson, G., & Katz, R. H. (1988). A Case for Redundant Arrays of Inexpensive Disks (RAID). SIGMOD '88: Proceedings of the 1988 ACM SIGMOD international conference on Management of data.

Section 3.3, "RAID 5: Rotated Parity Array": The paper describes RAID 5 as having high I/O rates for reads and large writes, with redundancy handled by parity. It notes the performance penalty for small writes: "a single-sector write involves 4 disk I/Os". This supports that it provides redundancy and some performance improvement but with a notable trade-off.

DOI: https://doi.org/10.1145/50202.50214

3. Stanford University, CS 140: Operating Systems Course Notes (Winter 2018):

Lecture 15: Filesystems, RAID, page 12: The notes describe RAID 10 (1+0) as "High performance (reads and writes), high availability." For RAID 5, it notes "Good for reads, not so good for writes (read-modify-write for parity)." This academic source clearly distinguishes the superior, well-rounded performance of RAID 10 while acknowledging that RAID 5 also offers a performance benefit, albeit a qualified one.