For a traditional SAN proposal, which element confirms scalable performance under growth?

A non-blocking fabric architecture ensures that any port can communicate with any other port on the same switch without internal contention, providing a high-performance foundation. Documenting a low oversubscription ratio (the ratio of end-port bandwidth to Inter-Switch Link bandwidth) of less than 1.5:1 is critical for scalability. It guarantees that as more devices are added, the core fabric has sufficient bandwidth to handle the aggregated traffic, preventing bottlenecks and ensuring predictable performance under growth. This design approach is a cornerstone of high-performance, scalable SANs.

B. Single HBA per host: A single HBA creates a performance bottleneck and a single point of failure. Scalable and resilient designs require dual HBAs for multipathing, load balancing, and high availability.
C. Daisy-chained switches: This topology is not scalable. It creates performance bottlenecks and increases latency as traffic from end switches must traverse intermediate switches, congesting the Inter-Switch Links (ISLs).
D. Only copper cabling: Limiting a design to only copper cabling severely restricts physical distance and scalability. Optical fiber is essential for connecting devices over longer distances, a common requirement in growing data centers.

1. HPE Storage SAN Design Reference Guide (April 2017):

Page 13, "Oversubscription": This section explains that "Oversubscription is the ratio of the sum of the bandwidths of all inputs to the sum of the bandwidths of all outputs." It further clarifies that managing this ratio is key to performance. A low ratio, as suggested in the correct answer, is indicative of a high-performance design.

Page 14, "Core-edge fabric": This section describes the core-edge topology as the most common for scalability and performance, explicitly contrasting it with less scalable options like daisy-chaining.

Page 10, "Host, storage, and switch considerations": This section emphasizes the need for redundant components, including HBAs, for high availability, which is a component of scalable enterprise solutions.

2. Brocade, "SAN Design Best Practices" Technical Brief: (HPE rebrands Brocade switches as B-series).

Section: "Fabric Design and Scalability": This document consistently emphasizes that a well-designed fabric with managed, low oversubscription is essential for predictable performance and scalability. It states that non-blocking switch architectures are a fundamental feature for enterprise SANs.

3. Cisco, "Cisco MDS 9000 Family Best Practices" Guide: (HPE rebrands Cisco switches as C-series).

Chapter: "Port-Channel and Virtual SAN Design Best Practices": This guide details the importance of ISL bandwidth and designing for specific oversubscription ratios to meet performance requirements. It notes that "A non-blocking architecture allows all ports to be active at line rate concurrently." This supports the principle that a non-blocking fabric is a key element for performance.