Bandwidth refers to the theoretical maximum data transfer capacity of a network link or medium, typically measured in bits per second (bps). It is an intrinsic property determined by the physical characteristics of the medium (e.g., fiber optic or copper cable) and the signaling technology used. While signal quality can degrade over extreme distances, requiring repeaters, the rated bandwidth of the link itself is not a function of its length. In contrast, delay, latency, and throughput are all directly or indirectly affected by the physical distance the signal must travel.

A. Delay: Propagation delay, a major component of total delay, is directly proportional to distance as it represents the time taken for a signal to travel across the medium.

B. Latency: Latency is the total time for a data packet to travel from source to destination. It includes propagation delay, which increases directly with distance.

D. Throughput: Throughput is the actual measured data transfer rate. It is often limited by latency over long distances, a phenomenon described by the bandwidth-delay product.

1. SNIA Dictionary. (2023). SNIA (Storage Networking Industry Association).

Bandwidth: Defined as "The data transfer capacity of a communications channel." This definition focuses on capacity, which is a property of the channel technology, not its length.

Latency: Defined as "The time it takes for a data packet to get from one designated point to another," which inherently includes the time to traverse a physical distance.

2. Kurose, J. F., & Ross, K. W. (2021). Computer Networking: A Top-Down Approach (8th ed.). Pearson.

Section 1.4.1, "Types of Delay in Packet-Switched Networks": The text explicitly defines propagation delay as "the distance between two routers divided by the propagation speed." This directly links delay and latency to physical distance. Bandwidth (referred to as transmission rate) is treated as a separate property of the link itself.

3. Cisco. (2006). WAN Performance Optimization. White Paper.

Page 2, "Bandwidth and Latency": This document distinguishes the concepts: "Bandwidth is the measure of the data-carrying capacity of a link... Latency is the time it takes for a packet to travel from its source to its destination." It further explains how latency, which increases with distance, negatively impacts throughput for protocols like TCP.

A dual-fabric design is the industry-standard architecture for achieving high availability (HA) and eliminating single points of failure (SPOF) in a storage network. In this model, two completely independent and redundant fabrics are created. Servers and storage devices are equipped with dual (or more) ports, with each port connecting to a separate fabric. If any single component—such as a host bus adapter (HBA), cable, switch, or even an entire fabric—fails, connectivity is maintained through the alternate path on the second fabric. This design also enhances performance and scalability, as traffic can be load-balanced across both fabrics using multipathing software, and each fabric can be scaled independently.

A. A large single fabric, regardless of its internal connectivity, represents a single point of failure. A fabric-wide service failure or misconfiguration would cause a complete outage.

C. High-quality components reduce the probability of failure but do not create a redundant architecture. A single path built with quality parts is still a single point of failure.

D. Clustered servers with single host adapters create a single point of failure at each server's connection to the SAN. The failure of that single adapter would isolate the server node.

1. IBM Redbooks. Introduction to Storage Area Networks, SG24-5470-06. (October 2006). Page 138, Section 4.5.2, "Redundant fabrics," states: "The most common way to provide high availability in a SAN is to use two completely separate fabrics... Servers and storage subsystems are attached to both fabrics. This is known as a dual-redundant fabric. If any component in one fabric fails, the servers can still access the storage subsystems through the other fabric."

2. Cisco Systems, Inc. Data Center SAN Design and Deployment Guide, Release 5.2. (February 2011). Chapter 2, "SAN Architecture," section "High Availability," details the dual-fabric model as the standard for production environments. It states, "The primary goal of a highly available design is to eliminate single points of failure... The recommended SAN design consists of a minimum of two independent fabrics."

3. Hewlett Packard Enterprise (HPE). HPE SAN Design Reference Guide. (April 2017). Page 11, section "High availability and resiliency," states: "The most fundamental design principle for a highly available SAN is the elimination of single points of failure (SPOFs). The best practice is to deploy a minimum of two identical and physically separate fabrics... Multi-ported hosts and storage systems connect to both fabrics."

This scenario requires balancing the Recovery Point Objective (RPO), Recovery Time Objective (RTO), and the time required to create a snapshot. The RPO of one hour dictates the maximum acceptable data loss, meaning a successful snapshot must be completed at least every hour. The RTO of 15 minutes requires that a consistent snapshot is always available to begin the recovery process immediately upon failure.

Since it takes 15 minutes to establish the snapshot, using a single snapshot volume would create a 15-minute window where no consistent, ready-to-use snapshot exists. Using two volumes in an alternating schedule solves this. While one volume is being updated, the other remains online and consistent, ensuring a recovery point is always available to meet the RTO. Scheduling the creation every 60 minutes aligns perfectly with the one-hour RPO, making it the most efficient solution without incurring the unnecessary overhead of more frequent snapshots.

A. Use one snapshot volume and incrementally establish it every 60 minutes.

A failure during the 15-minute snapshot creation would make the previous snapshot 75 minutes old, violating the one-hour RPO.

B. Use one snapshot volume and incrementally establish it every 30 minutes.

For 15 minutes of every 30-minute cycle, the single snapshot volume is being updated and is unavailable for an immediate restore, failing to guarantee the RTO.

D. Use two snapshot volumes and incrementally establish one volume every 30 minutes.

This meets the RPO and RTO but is inefficient. It creates snapshots twice as frequently as required by the one-hour RPO, consuming unnecessary system and network resources.

1. Somasundaram, G., & Shrivastava, A. (Eds.). (2012). Information Storage and Management: Storing, Managing, and Protecting Digital Information in Classic, Virtualized, and Cloud Environments (2nd ed.). Wiley.

Chapter 11, Section 11.2.1, "BC Terminology": Defines RPO as "the point in time to which the systems and data must be recovered... RPO is generally defined in terms of hours." This supports matching the snapshot frequency to the one-hour RPO for efficiency.

Chapter 9, Section 9.3.2, "Uses of Local Replicas": Explains that replicas are used for recovery from data corruption. It emphasizes the need for a consistent, point-in-time copy. The logic of using multiple replicas to ensure one is always consistent and available while another is synchronizing is a direct application of this principle, especially when the synchronization time (15 minutes) is significant.

2. Gibson, G. A. (1996). CS-96-162 Course Syllabus: 15-740 Computer Architecture. Carnegie Mellon University, School of Computer Science.

In course materials discussing storage systems and reliability, the fundamental concept is that a recovery mechanism must have a consistent, available data copy to restore from. The time to create this copy (the 15-minute snapshot establishment) represents a window of vulnerability if only one copy exists, directly impacting the ability to meet a strict RTO. The use of alternating buffers (i.e., two snapshot volumes) is a standard technique to mitigate this.

The detailed design phase of an Information Lifecycle Management (ILM) project focuses on creating a specific, actionable blueprint for implementation. This phase translates the high-level, conceptual design into precise technical specifications and project plans. Defining internal storage parameters (e.g., RAID levels, tiering policies, LUN configuration) is a core technical task of this phase. Concurrently, preparing a work-task breakdown structure is a critical project management activity that outlines all the specific tasks, dependencies, and resources required for the implementation, ensuring a structured and manageable deployment process.

A. Developing an idealized environment is a characteristic of the high-level or conceptual design phase, where the overall target architecture is defined without specific implementation details.

D. This activity is typically performed during the initial analysis or high-level design phase to justify the project and outline the strategic benefits and architectural changes.

1. Somasundaram, G., & Shrivastava, A. (Eds.). (2012). Information Storage and Management: Storing, Managing, and Protecting Digital Information in Classic, Virtualized, and Cloud Environments (2nd ed.). Wiley.

Reference: Chapter 13, Section "Storage Design and Implementation," subsection "Physical Design Phase." This section details that the physical design (detailed design) phase involves creating the implementation plan and defining specific configuration details for storage, which directly supports defining parameters (C) and preparing a work breakdown structure (B).

2. SNIA. (2004). Information Lifecycle Management (ILM) Initiative. SNIA Technology Center White Paper.

Reference: The general principles outlined in SNIA's ILM initiative describe the progression from high-level policy creation to the detailed implementation of those policies on storage infrastructure. The detailed design phase is where abstract policies are mapped to concrete storage parameters and implementation tasks.

3. MIT OpenCourseWare. (2004). 16.852J / ESD.34J System Engineering, Lecture 6: Work Breakdown Structure (WBS). Massachusetts Institute of Technology.

Reference: Slide 4, "WBS Definition." This courseware defines the WBS as a tool to break down a project into manageable work packages. This is a fundamental activity in the detailed planning and design stage before implementation begins.

The scenario describes a need for multiple users to access shared files (manuals, drawings) that are relatively static but frequently used. A Network Attached Storage (NAS) system is the ideal solution as it is specifically designed for file-level access over a standard Ethernet network using protocols like NFS or SMB/CIFS. This approach provides a centralized, reliable, and reasonably fast method for sharing files without the complexity and cost of a block-level Storage Area Network (SAN). Creating a dedicated Ethernet segment for the NAS can further ensure performance by isolating storage traffic. This directly matches the need to serve files to many users.

A. FC-Arbitrated Loop is block-level storage. Access via a single server creates a bottleneck and is not a direct method for sharing files among many users.

B. Direct block-level connections to all servers require a complex clustered file system for safe file sharing, which is not a simple or standard solution for this use case.

D. A Fibre Channel SAN provides block-level access and is overly complex and expensive for a simple file-sharing requirement. It does not natively serve files to users.

---

1. SNIA (Storage Networking Industry Association). SNIA Dictionary.

Definition of Network Attached Storage (NAS): "A term used to refer to storage devices that connect to a network and provide file access services to computer systems. These devices usually attach to an existing Ethernet network." This supports option C as the most direct method for providing file access services.

Definition of Storage Area Network (SAN): "A network whose primary purpose is the transfer of data between computer systems and storage elements... A SAN provides block-level access to storage." This clarifies why options B and D, which describe block-level SAN/direct-connect architectures, are inappropriate for a direct file-sharing need.

Reference Location: Accessible via the official SNIA website's educational resources section.

2. University of California, Berkeley. EECS Department. CS 162: Operating Systems and System Programming, Lecture 20.

Content: The lecture notes explicitly contrast NAS and SAN. It states, "NAS: Provides a file system interface (NFS, CIFS)... Clients see a file system, not raw blocks." and for SAN, "SAN: Provides block-level interface... Clients see a raw disk." This academic source confirms that for serving files, a NAS is the appropriate architecture.

Reference Location: Slide 43, "NAS vs. SAN," in the "Filesystems 3 (Cloud, Distributed)" lecture presentation.

3. SNIA Education. "Storage Networking Foundations" Tutorial.

Content: This foundational SNIA tutorial explains the fundamental differences between storage architectures. In the section comparing NAS and SAN, it highlights that NAS is optimized for file sharing and collaboration over standard IP networks, making it easy to deploy for use cases like the one described. It positions SAN as a solution for high-performance block I/O, typically for databases and application servers, which is overkill and architecturally incorrect for this scenario.

Reference Location: Section titled "Network Attached Storage (NAS)" and "Storage Area Networks (SAN)".

The requirement for "no single point of failure" (SPOF) mandates a redundant design. For 32 servers to access storage without a SPOF, each server requires at least two independent paths. This is typically achieved by installing two Host Bus Adapters (HBAs) in each server, connecting to separate fabrics. This results in a total of 32 servers × 2 HBAs/server = 64 HBAs.

The disk subsystem's total throughput is 8 ports × 200 MB/s per port = 1600 MB/s. This aggregate bandwidth must be shared among all active connections. Dividing the total throughput by the total number of HBAs gives the average throughput per HBA: 1600 MB/s ÷ 64 HBAs = 25 MB/s.

A. 16 HBAs is insufficient for 32 servers, failing both the access and redundancy requirements.

B. 32 HBAs provides only a single path per server, which creates a single point of failure, violating the design requirement.

D. The required throughput (64 HBAs × 50 MB/s = 3200 MB/s) would exceed the storage subsystem's maximum capacity of 1600 MB/s.

1. Farley, M. (2004). Storage Networking Fundamentals. Cisco Press. Chapter 4, "High Availability and Business Continuance," in the section "Designing Redundancy in the I/O Path," explains that eliminating single points of failure requires redundant components, including at least two HBAs per server.

2. Livny, M. (2005). Introduction to Storage Area Networks (SANs) [Presentation for CS 736, University of Wisconsin-Madison]. Slide 21, "High Availability," graphically depicts a fully redundant SAN architecture where servers have dual HBAs to connect to separate fabrics, ensuring no single point of failure.

3. SNIA (Storage Networking Industry Association). SNIA Dictionary. The definition for "Multipathing" states it is "The use of more than one path for I/O between a host computer and a storage device... Multipathing is a fault tolerance technique..." This principle justifies the use of 64 HBAs (two per server) to create multiple, redundant paths.

The question asks for the Maximum Transmit Unit (MTU) size for IP-based storage. While the term MTU typically refers to the maximum size of a data packet that a link-layer protocol like Ethernet can carry (e.g., 9000 bytes for jumbo frames), this is not among the options. The question is referring to the maximum theoretical size of a datagram at the IP layer. The IPv4 packet header contains a 16-bit "Total Length" field. This field specifies the total datagram size in bytes, including the header and data. A 16-bit field allows for a maximum value of 2^16 - 1, which equals 65,535 bytes (approximately 65K bytes). This is the largest possible logical packet the IP protocol can handle.

A. 2K bytes: This is a small, non-standard jumbo frame size and does not represent the maximum limit of the IP protocol.

C. 1508 bytes: This value is very close to the standard Ethernet MTU of 1500 bytes, not the size for jumbo frames or the maximum IP datagram size.

D. 100M bytes: This size is excessively large and far exceeds the 16-bit length field limitation of the IPv4 protocol.

---

1. Postel, J. (1981). RFC 791: Internet Protocol. Defense Advanced Research Projects Agency. Section 3.1, Page 11.

This foundational document for the Internet Protocol specifies the packet format. It states: "Total Length: 16 bits. Total Length is the length of the datagram, measured in octets, including internet header and data... This field allows the length of a datagram to be up to 65,535 octets."

2. Kurose, J. F., & Ross, K. W. (2017). Computer Networking: A Top-Down Approach (7th ed.). Pearson. Chapter 4: The Network Layer: Data Plane, Section 4.4.1: The IPv4 Datagram Format.

This standard university textbook explains that the 16-bit "datagram length" field in the IPv4 header limits the total size of an IP datagram to 65,535 bytes. This is a core concept taught in reputable computer science and networking courses.

3. Stanford University. (2013). CS144: Introduction to Computer Networking, Lecture 7: IP Forwarding. Slide 11, "IP Header Fields".

The course material explicitly shows the 16-bit "Total Length" field in the IPv4 header and notes the "Max size = 2^16 - 1 = 65535 bytes". This confirms the theoretical maximum packet size as a fundamental part of IP taught in higher education. (Available at: https://web.stanford.edu/class/cs144/lectures/07-ip-forwarding.pdf)

According to ITIL principles, the lifecycle of addressing an error extends beyond immediate remediation. The process is divided into Incident Management (restoring service) and Problem Management (preventing recurrence). A storage administrator's job is not complete after fixing the immediate issue (the incident). The role, under the Problem Management process, requires performing a root cause analysis to understand why the error occurred and then implementing a permanent solution or change to ensure it does not happen again. Therefore, the job is truly complete only when a plan is developed and implemented to prevent future occurrences.

A. When the problem is fixed: This describes the completion of Incident Management, which focuses on restoring service quickly, not the completion of the entire Problem Management process.

B. When the cause of the error is determined: This is a critical intermediate step (root cause analysis) in Problem Management, but it is not the final action; a solution must still be implemented.

C. When the problem is identified and the vendor is notified: This is an early diagnostic or escalation step in the process, not the resolution or conclusion of the administrator's responsibility.

---

1. University of Texas at Austin, Information Technology Services. ITIL Problem Management Process Guide. This guide outlines the full lifecycle. The final steps, "Problem Resolution & Closure," involve implementing a permanent fix via a Request for Change (RFC) to prevent the problem from recurring. The process is closed only after verifying the fix is successful. (See Section 4.8, "Problem Resolution & Closure").

2. University of Washington, IT Connect. ITIL V3 Foundation Courseware. This educational material defines the primary objective of Problem Management as: "to prevent Problems and resulting Incidents from happening, to eliminate recurring Incidents and to minimize the impact of Incidents that cannot be prevented." This directly supports the idea that preventing recurrence is the end goal. (See Module: "Problem Management," Section: "Goals & Objectives").

3. EMC Education Services. Information Storage and Management: Storing, Managing, and Protecting Digital Information in Classic, Virtualized, and Cloud Environments, 2nd Edition. This textbook, a foundational resource for SNIA certifications, discusses storage management best practices. Chapter 15, "Storage Infrastructure Management," emphasizes proactive management activities, which align with the ITIL goal of preventing problems rather than just reacting to them. (See Chapter 15, Section: "Monitoring the Storage Infrastructure").

LUN masking is a mandatory access control feature implemented on the storage array controller. It functions by creating an access control policy that maps specific Logical Unit Numbers (LUNs) to the unique World Wide Port Names (WWPNs) of the server Host Bus Adapters (HBAs). When an HBA logs into the storage port, the array consults this policy and only presents (makes visible) the LUNs that are explicitly mapped to that HBA's WWPN. This is the industry-standard method for enforcing granular, LUN-level access control from the target side, ensuring a server cannot discover or access LUNs not assigned to it.

A. Configure the HBA for persistent binding: This ensures a LUN is always assigned the same device name in the OS across reboots; it does not control or prevent access.

B. Disable the undesired LUNs in the operating system disk manager of each server: This is a weak, host-based configuration, not a security mechanism. A privileged user on the server could easily re-enable the LUNs.

D. Substitute the hub with a switch and use soft zoning to assign RAID LUNs to HBA WWNs in the servers: Zoning is a switch-level function (the scenario uses a hub) that controls device-to-device communication, not LUN-level access.

1. Storage Networking Industry Association (SNIA), SNIA Dictionary, "LUN Masking". The official definition states: "A process that makes a LUN available to some hosts and unavailable to other hosts." This directly supports the function described in the correct answer.

(Available at: https://www.snia.org/education/online-dictionary)

2. Menon, J., & IBM Almaden Research Center. (2003). Storage Networking: Technology, Architectures, and Management. In Communications of the ACM, 46(7), pp. 81-88. This publication discusses SAN security principles, noting that storage controllers implement access control mechanisms like LUN masking to restrict host access to specific storage volumes. Section: "Security and Management".

DOI: https://doi.org/10.1145/792704.792707

3. Gibson, G. A., & Van Meter, R. (2000). Network attached storage architecture. In Communications of the ACM, 43(11), pp. 37-45. While focused on NAS, this foundational paper discusses storage access control principles. The concept of restricting access at the storage device level is fundamental and analogous to LUN masking in a SAN, which restricts access based on the initiator's identity (WWPN).

DOI: https://doi.org/10.1145/354320.354337

The question centers on the fundamental decision between block-level and file-level storage access when consolidating servers.

Block-level access is provided by a Storage Area Network (SAN). The server's operating system sees the storage as a raw, unformatted disk volume. This is required for applications that need to control the file system directly, such as database servers or hypervisors (e.g., VMware's VMFS). A traditional Network Attached Storage (NAS) device, which serves files and not blocks, cannot fulfill this requirement.

File-level access is provided by a NAS. The storage is presented to servers as a network file share (e.g., via NFS or SMB protocols). If applications only need to read and write to a pre-existing file system, NAS is a suitable and often more economical and simpler solution to implement and manage than a SAN, as it leverages standard Ethernet networking.

A. While modern NAS can offer high throughput, SANs are typically chosen for latency-sensitive, high-performance transactional workloads. The term "high throughput" is too general to make this a universally true statement for selection.

D. This statement focuses on a secondary requirement (backups) rather than the primary access method (block vs. file), which is the core decision point mentioned in the question's premise.

1. SNIA (Storage Networking Industry Association). SNIA Dictionary.

Definition of Network Attached Storage (NAS): "A term used to refer to storage devices that connect to a network and provide file access services to computer systems." This supports the reasoning that NAS is file-level.

Definition of Storage Area Network (SAN): "A network whose primary purpose is the transfer of data between computer systems and storage elements... The term SAN is usually (but not necessarily) identified with block I/O services." This supports the reasoning that SAN is for block-level access, making NAS unacceptable if block is required (Answer B).

2. Prasad, R., & S. R. (2021). Storage Networks and Data Protection. Wiley.

Chapter 2, Section 2.3 "Network-Attached Storage (NAS)": This chapter explains that NAS systems are specialized file servers that are easy to install and manage, often leveraging existing IP networks, which contributes to their cost-effectiveness compared to dedicated Fibre Channel SANs. This supports the economic argument in Answer C.

Chapter 3, Section 3.2 "Block-Level Access in SAN": This section details how SANs present storage as block devices, which is a prerequisite for servers to create and manage their own file systems (e.g., NTFS, VMFS). This reinforces why a file-serving NAS is unacceptable when block access is a requirement (Answer B).

3. Tanenbaum, A. S., & Van Steen, M. (2017). Distributed Systems: Principles and Paradigms (3rd ed.).

Chapter 11, "Distributed File Systems": This chapter discusses architectures like NFS, which are fundamental to NAS. It contrasts this with block-level services where the client machine's OS is responsible for the file system organization. This academic text confirms the fundamental architectural difference between file-level (NAS) and block-level (SAN) access. This distinction is the basis for both correct answers B and C.