Scribd
Upload
0 views5 pages

dstn1

The document discusses various aspects of HDD performance measurement, emphasizing the importance of IOPS over MBps for random workloads. It also compares different HDDs for performance, evaluates RAID setups for capacity utilization, and addresses issues with NFS versions and SCSI-based storage implementation. Additionally, it covers the impact of fragmentation on HDDs and SSDs, the benefits of LBA for disk addressing, and proposes hybrid storage solutions for application loading and dynamic storage needs.

Uploaded by

yadavshivjee18
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
0 views5 pages

dstn1

The document discusses various aspects of HDD performance measurement, emphasizing the importance of IOPS over MBps for random workloads. It also compares different HDDs for performance, evaluates RAID setups for capacity utilization, and addresses issues with NFS versions and SCSI-based storage implementation. Additionally, it covers the impact of fragmentation on HDDs and SSDs, the benefits of LBA for disk addressing, and proposes hybrid storage solutions for application loading and dynamic storage needs.

Uploaded by

yadavshivjee18
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
You are on page 1/ 5

Q1. Why is IOPS preferred over MBps for HDD performance measurement?

Answer:

IOPS (Input/Output Operations Per Second) measures the number of read/write operations, which
matters most for random workloads.

MBps (Megabytes per second) measures throughput (data transfer speed), which is useful for sequential
workloads.

HDDs have mechanical parts (spinning platters and moving heads), so their real-world performance
depends on how fast they can process I/O requests rather than raw data transfer speed.

👉 IOPS is more relevant because HDDs handle random requests with varying speeds.

Q2. Which HDD is better for random workloads?

Answer:

Given two options:

2.5-inch, 900GB, 15K RPM, 3.9ms seek time, 198MBps

3.5-inch, 4TB, 10K RPM, 5.4ms seek time, 200MBps

👉 Choose the 2.5-inch, 15K RPM HDD because:

✅ Faster seek time (3.9ms vs. 5.4ms) → Better for random access.

✅ Higher RPM (15K vs. 10K) → Faster rotations = quicker data retrieval.

✅ IOPS is more critical than MBps for random workloads.

Q3. Which RAID setup has the best capacity utilization?

Options:

RAID 5 (7+1)

RAID 6 (30+2)

RAID 6 (6+2)

Answer:

👉 RAID 5 (7+1) is the most efficient in terms of capacity utilization.


RAID 5 uses (N-1) disks for data storage, meaning 87.5% usable capacity (7/8 disks).

RAID 6 uses (N-2) disks for data, meaning:

30+2: ~93.75% utilization (30/32)

6+2: 75% utilization (6/8)

📌 Conclusion: RAID 5 (7+1) provides a good balance of storage and redundancy.

Q4. Which NFS version has firewall issues?

Answer:

👉 NFS v2 & v3 have firewall issues because they use multiple ports dynamically.

✅ NFS v4 solves this problem by using a single TCP port (2049).

📌 Conclusion: NFS v4 is easier to configure with firewalls.

Q5. How should the company implement SCSI-based storage?

Answer:

Since the company has no existing storage but a strong network, they should use:

✅ iSCSI (Internet SCSI) → Uses TCP/IP to connect storage over a network.

✅ Justification:

Cost-effective (uses Ethernet instead of expensive Fibre Channel).

Provides block-level storage.

Scales well for future storage expansion.

📌 Conclusion: iSCSI is the best choice due to flexibility and performance over the existing network.

Q6. How long does it take to read 2000 sectors from a disk?

Given:

Seek Time = 4ms

Rotational Speed = 10,000 RPM

Rotational Latency = Half of full rotation

Sectors per Track = 600


Each sector = 512 bytes

Total Sectors to Read = 2000

Solution:

👉 Time per sector = Seek Time + Rotational Latency

Rotational Latency = (60,000 ms / 10,000 RPM) / 2 = 3ms

Total time per sector = 4ms + 3ms = 7ms

Total Time for 2000 sectors = 2000 × 7ms = 14,000ms = 14 seconds

📌 Final Answer: 14,000 milliseconds (14 sec).

Q7. RAID 4: What is the missing disk data?

Given data from 3 disks:

1100

1010

1111

👉 RAID 4 uses Parity (XOR) to recover missing data

XOR Rule: A ⊕ B ⊕ C ⊕ D = 0

Missing Disk Data:

1100 ⊕ 1010 ⊕ 1111 = 0001

📌 Final Answer: 0001

Q8. Cost-effective solution for slow application loading and dynamic storage needs

Problem:

Slow application loading (10GB files need fast access).

Dynamic 1TB storage required.

Solution:
✅ Use SSD for fast application loading (better read speeds than HDD).

✅ Use HDD for 1TB dynamic storage (cheaper, high capacity).

✅ Hybrid Approach (SSHD or Tiered Storage) to balance cost and performance.

📌 Conclusion: Use SSD for OS & application files, HDD for data storage.

Q9. Do file fragments cause performance issues? How does it differ for HDD vs. SSD?

Answer:

👉 Yes, fragmentation causes performance issues, but differently for HDD & SSD.

For HDD:

❌ Bad performance → Read/write head moves a lot.

✅ Solution: Defragmentation tools.

For SSD:

✅ Minimal impact → No moving parts.

❌ Defragmentation is harmful → Reduces SSD lifespan.

📌 Conclusion: Fragmentation affects HDDs more than SSDs.

Q10. Why is disk addressing abstracted by LBA?

Answer:

Problem: Directly accessing disk blocks (CHS - Cylinder, Head, Sector) is complex and inefficient.

Solution:

👉 LBA (Logical Block Addressing) abstracts physical disk structure and assigns a unique number to each
block.

Diagram Explanation:
mathematica

Copy

Edit

Traditional Addressing → CHS (Cylinder, Head, Sector)

LBA Addressing → Linear Block Number (0,1,2,3,...)

✅ Benefits:

Simplifies OS and software interaction with storage.

Works across different drive technologies (HDD, SSD).

Improves scalability.

📌 Conclusion: LBA makes disk management easier and faster.

You might also like