Pith. sign in

REVIEW 3 major objections 5 minor 21 references

An Optimized Disk Scheduling Algorithm With Bad-Sector Management

T0 review · 3 major / 5 minor · reviewed 2026-08-14 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read A LOOK-style scheduler that groups requests by cylinder cuts total disk access time by 33.53% and manages bad sectors.

desk verdict Minor LOOK variant with real but small novelty; the reported gains are arithmetic from a self-defined cost model and the bad-sector mechanism contradicts its own definition. read the letter →

arxiv 1908.01167 v1 pith:CS75EYAV submitted 2019-08-03 cs.OS cs.DS

Amar Ranjan Dash , Sandipta Kumar Sahu , B Kewal This is my paper · ORCID
classification cs.OScs.DS
keywords DiskschedulingBad-sectormanagementSeektimeRotationallatencyDatatransferMulti-platterhardLOOKalgorithmOperatingsystems
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper proposes MODSBSM, a disk scheduler that first sorts the request queue by track, then decides the arm's direction by comparing the distance to the lowest and highest requested tracks. When the head reaches a track, the scheduler completes every request on that whole cylinder—ordering sectors and platters directionally—before moving on, so it attacks rotational latency and head-switch transfer time rather than seek time alone. Across six twenty-request workloads on single- and multi-platter disks, the paper reports a 33.53% reduction in total disk access time relative to six traditional algorithms and a 7.51% reduction relative to five recent seek-time-optimized LOOK variants. The same pass maintains a bad-sector list: an index that fails twice is recorded, and later reads use a stored preferred bit, which the authors estimate saves n−2 retries per bad index in power and heat. The intended contribution is a single ordering discipline plus a defect-handling table that treats the hard disk as a multi-platter physical device.

What carries the argument

The load-bearing object is a physical-address ordering of the request queue: a linked list where each node holds platter, track, sector, read/write flag, and a bad-sector index (BSI). The list is sorted by track, and the arm chooses a direction by comparing LD, the distance from the initial head track to the first request's track, with RD, the distance to the last request's track. Whichever direction is chosen, the scheduler batches all requests on one track and cylinder, ordering them by sector and then platter so the actuator's head stack serves the entire cylinder while it is in position. Rotational latency is charged as the signed sector difference on an eight-sector track, and data transfer time as the platter difference plus one, so the ordering directly determines the reported savings. A separate Bad-Sector List records indices whose BSI reaches two; the BSM function then reads or writes using a stored preferred bit, flipping it once if the first guess fails.

What would settle it

A concrete falsifier: run the same six request patterns on a real hard disk (or a validated drive simulator with actual seek curves, rotational position, and transfer rates) and compare MODSBSM's total I/O time against LOOK, C-LOOK, and the five derived algorithms; the central claim fails if the total access time does not come out roughly one-third below the traditional set and 7.51% below the derived set. A second test: take a genuinely defective physical sector, attempt two reads, and see whether a stored preferred bit yields correct data where ECC and remapping would normally be required.

Watch

Extended reading notes

Core claim

The central claim is that optimizing only seek time, as traditional and recent LOOK-style algorithms do, leaves rotational latency and data transfer time on the table; a scheduler that also orders requests within a cylinder can reduce all three. After sorting requests by track and choosing the nearer end of the request range, MODSBSM visits each cylinder once, reading its sectors in ascending or descending order and its platters in the matching order, so sector gaps and platter-to-platter head switches are counted exactly under the paper's model. The paper's six cases, each with twenty requests, give MODSBSM a total disk access time of 1886 units versus 2228 for the best traditional LOOK and 1984 for the best recent derived variant; it phrases this as a 33.53% improvement over traditional algorithms and 7.51% over five derived algorithms. The algorithm's second, independent claim is that a sector whose bad-sector index reaches two is moved into a bad-sector table and resolved by storing a preferred bit, so the drive does not retry the same failed location repeatedly.

Load-bearing premise

The reported gains rest on a synthetic cost model in which seek, rotational latency, and transfer times are unit costs (one per track, one per sector gap on an eight-sector track, one per platter difference plus one), and on the assumption that a twice-failed sector can be recovered by storing and flipping a preferred bit; if real drive timings or real defect behavior differ, the 33.53% and 7.51% margins could shrink or disappear.

Editorial extensions

If this is right

  • If the ordering result holds on real hardware, a disk scheduler can reduce all three access-time components at once by serving every request on a cylinder before moving the arm, not just by minimizing head travel.
  • The reported figures translate the rule into numbers: 33.53% lower total disk access time than the six traditional algorithms and 7.51% lower than the five recent LOOK-style variants on the paper's six single- and multi-platter workloads.
  • With a bad-sector list, a sector that fails twice is recorded with a preferred bit, and later reads skip repeated retries; the paper estimates this removes n−2 unnecessary reads per bad index, lowering both power draw and heat.
  • Because the actuator arm moves all read/write heads together, the cylinder-first traversal matches the physical geometry of multi-platter drives, which the paper argues earlier single-platter-oriented algorithms ignore.
  • The algorithm's per-request structure (platter, track, sector, read/write flag, bad-sector index) gives the operating system a concrete data structure for scheduling and defect handling in one pass.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • The paper's cost model uses an eight-sector track and unit transfer times; with zone bit recording, on-drive caches, and native command queuing, the sector-gap ordering's advantage is likely smaller than the reported 7.51% margin.
  • The bad-sector preferred-bit mechanism can be tested apart from scheduling: if sectors with real media defects are already remapped by the drive firmware, the scheduler's bad-sector list would mostly see uncorrectable ECC errors, and flipping a preferred bit would not recover data.
  • A natural transfer of the idea is to SSDs or NVMe, where the equivalent of cylinder locality is logical-block and page locality and the costs are erase/write latency and wear; the paper only lists SSDs as future work, so this is an extension rather than a claim.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 5 minor

Summary. The manuscript proposes MODSBSM (Modern Optimized Disk Scheduling with Bad-Sector Management), a LOOK-style scheduler that orders requests by track, then sector, then platter, and includes a 'bad sector list' with a prescribed-bit recovery mechanism. The authors evaluate MODSBSM against six traditional algorithms and five recent LOOK-style algorithms on six hand-constructed 20-request cases covering single- and multi-platter disks with ascending, descending, and random track orders. They report a 33.53% reduction in total disk access time relative to traditional algorithms and a 7.51% reduction relative to the referred algorithms (Section 4.8, Table 3), and claim the algorithm detects and resolves bad sectors while reducing power consumption and heat. The paper's central assertions are that MODSBSM performs better than all compared algorithms under the stated performance measure and that its bad-sector management is functional.

Significance. If the performance claims were supported by realistic evaluation, MODSBSM would be a useful contribution as a LOOK-style scheduler that explicitly accounts for platter geometry and sector ordering. The paper is also commendable for presenting a complete pseudocode and a defined cost model that allows reproduction of the reported tables. However, the evaluation is entirely based on a self-defined cost model with no real disk measurements or trace replay, and the bad-sector mechanism is internally inconsistent with the paper's own definition of a bad sector. As a result, the central claims are not established. The work may serve as a thought experiment, but it is not yet a validated disk scheduling algorithm.

major comments (3)
  1. [Section 4.1 and Steps 36-49, 87-96] The evaluation rests on a cost model in which rotational latency is the sector-number difference on an 8-sector track (Steps 36-39, 45-48), transfer time is the absolute platter-number difference plus a unit transfer time (Steps 40, 49, 87, 96), and seek time is simply the track difference. These unit costs are not tied to any real disk timing; real drives have nonlinear seek curves, nonzero head-switch and settle times, zone-based recording, and command queuing. Because MODSBSM is constructed to minimize exactly these quantities by sorting sectors and platters, its advantage under this model follows by construction. The 33.53% and 7.51% improvements in Section 4.8 are arithmetic ratios of these invented units and have no demonstrated external validity. To make the claim load-bearing, the authors would need to validate the cost model against real disk timings or a calibrated simulator and confirm the ranking on realistic trace workloads.
  2. [Section 3.1 Steps 22-31 and 69-78, BSM() Steps 4-8, Section 5] The algorithm defines a bad sector as 'unreadable and non-writable' (Steps 23 and 70), yet BSM() resolves it by flipping a 'preferred bit' and then performing the memory operation (Steps 4-8). No inversion of a bit can recover data from a physically unreadable medium, and the algorithm never remaps the sector to a spare area, which is how real drives handle bad sectors. The energy and heat savings formulas in Section 4.8 ('e*n-2' and 'h*n-2') are asserted without derivation or measurement, and the notation is undefined. These issues invalidate the claim in Section 5 that the algorithm can detect and resolve bad sectors.
  3. [Sections 4.2-4.7 and Table 3] The comparative analysis uses only six hand-constructed 20-request cases with no sensitivity analysis, varying request counts, or random sampling. The totals in Table 3 are sums over these six cases; for example the MODSBSM total disk access time is 1886 versus 1984 for SMCC, a roughly 5% difference that could easily reverse under a different but realistic unit-cost weighting. Without statistical analysis or workload diversity, the conclusion in Section 4.8 that the algorithm 'performs better than all traditional and latest modified disk scheduling algorithms' is an overgeneralization.
minor comments (5)
  1. [Section 3.1, Steps 2, 15, 52, 62, 100] The variable TSKT is initialized in Step 2, but the algorithm later updates 'SKT' in Steps 15, 52, 62, and 100; this appears to be an inconsistency and should be fixed by using TSKT throughout.
  2. [Section 4.6] The caption for the case-5 figure reads 'Figure 2: Working of all Disk-Scheduling algorithms for Case-5'; this should likely be Figure 16.
  3. [Section 4.7] Section 4.7 contains a typo: 'Figure 20represents' should be 'Figure 20 represents'.
  4. [Section 3.2] In Section 3.2, 'tack' appears instead of 'track' in the phrase 'with in each tack-sector combination'.
  5. [Abstract] The abstract states 'researchers try to optimize the CPU Scheduling algorithms' but the paper addresses disk scheduling; the opening sentences should be clarified to avoid confusion.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the performance comparison is computed under the paper's stated cost model, and no fitted parameter, self-citation chain, or imported uniqueness theorem drives the central claim.

full rationale

The paper claims MODSBSM reduces total disk access time by 33.53% and 7.51%, but this is not a 'prediction' in the circularity sense. Section 4.1 explicitly fixes the performance measure: seek time is head movement (track distance), rotational latency is sector distance on an 8-sector track, transfer time is |platter difference| + 1, and disk access time is the sum. The algorithm's Steps 11-13 and 58-60 sort requests by track, then sector, then platter, which is a reasonable construction for minimizing those exact quantities, but the comparison in Table 3 is an arithmetic evaluation of that model, not a parameter fitted to the claims. No parameter is calibrated to the target improvement; the claimed percentages are just ratios of the computed totals. The bad-sector 'resolution' by flipping a preferred bit is physically inconsistent with the paper's own definition of a bad sector as unreadable/non-writable (Steps 22-23, 69-70), and the energy/heat formulas (e*n-2, h*n-2) are definitional restatements rather than measurements; however, inconsistency and lack of external validation are correctness/threats-to-validity issues, not circularity. The paper's self-citations [4], [5] concern CPU scheduling and are not load-bearing for the disk-scheduling result. No uniqueness theorem is invoked. Therefore the derivation chain does not reduce to its inputs; score 0.

Assumptions & free parameters 3 free parameters · 5 assumptions · 2 invented entities

The central performance claim depends on a custom cost model (unit transfer time, platter-difference head selection, 8-sector wrap-around), a zero-sorting-time assumption, the single-actuator cylinder assumption, and an unvalidated bad-sector bit-inversion mechanism. The free parameters are hand-chosen unit costs and the detection threshold of two failures. The invented bad-sector list and BSM() function have no independent evidence.

free parameters (3)
  • unit data transfer time = 1
    Section 4.1 states 'data transfer (assumed as one)' for each request; this unit cost is chosen by hand and enters every TDAT comparison.
  • per-platter head selection cost = 1 per platter difference
    TDTT is computed as the absolute difference in platter numbers (Steps 40, 49, 87, 96), treating each platter switch as one time unit; no hardware basis is given.
  • bad sector detection threshold = 2
    A memory index is declared bad only after it fails twice (BSI reaches 2); this threshold is arbitrary and not derived from error rates.
assumptions (5)
  • domain assumption All read/write heads are attached to a single actuator and lie on the same cylinder.
    Section 3.2 and Section 2 state this hardware property; used to justify serving all requests in a cylinder without seeking.
  • domain assumption Each track has exactly 8 sectors numbered 0 through 7.
    Rotational latency formulas in Steps 36-39 and 45-48 use 7 as the maximum sector index; no disk geometry is cited.
  • domain assumption Sorting time of the request queue is zero.
    Section 4.1: 'The time required for sorting of memory requests assumed as zero.' This affects all algorithms equally but simplifies the comparison.
  • ad hoc to paper Bad sectors can be modeled as invertible bit flips fixed by a preferred bit.
    Section 3.2 describes incrementing BSI on failed reads and later using a preferred bit to finalize reads; physical bad sectors are not corrected by flipping a preferred bit.
  • domain assumption Sector ordering within a cylinder minimizes rotational latency.
    The within-track sector sorting in Steps 12-13 and 58-60 assumes that processing sectors in ascending order minimizes total rotation; this ignores the disk's continuous rotation and head-switching delays.
invented entities (2)
  • Bad Sector List with preferred-bit column
    purpose: Track sectors that failed twice and store a preferred bit (0 or 1) to use in subsequent reads.
    The paper provides no hardware validation or external test showing this mechanism corrects real bad sectors; it is a data structure defined in Section 3 and Table 2.
  • BSM() management function
    purpose: Resolve bad sectors by checking the finalized flag and prescribed bit before performing the memory operation.
    Described in Section 3.1; no independent implementation or measurement of its effectiveness is provided.

how reviews work

0 comments
Cite this review

Pith. "Pith review of An Optimized Disk Scheduling Algorithm With Bad-Sector Management." pith.science (2026). https://pith.science/paper/CS75EYAV

@misc{pith2026190801167,
  author       = {Pith},
  title        = {Pith review of: An Optimized Disk Scheduling Algorithm With Bad-Sector Management},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/CS75EYAV}},
  note         = {Machine review of arXiv:1908.01167}
}
read the original abstract

In high performance computing, researchers try to optimize the CPU Scheduling algorithms, for faster and efficient working of computers. But a process needs both CPU bound and I/O bound for completion of its execution. With modernization of computers the speed of processor, hard-disk, and I/O devices increases gradually. Still the data access speed of hard-disk is much less than the speed of the processor. So when processor receives a data from secondary memory it executes immediately and again it have to wait for receiving another data. So the slowness of the hard-disk becomes a bottleneck in the performance of processor. Researchers try to develop and optimize the traditional disk scheduling algorithms for faster data transfer to and from secondary data storage devices. In this paper we try to evolve an optimized scheduling algorithm by reducing the seek time, the rotational latency, and the data transfer time in runtime. This algorithm has the feature to manage the bad-sectors of the hard-disk. It also attempts to reduce power consumption and heat reduction by minimizing bad sector reading time.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

21 extracted references · 20 canonical work pages

  1. [1]

    Galvin, and Gerg Gagne, (2014) ―Operating System Concepts ‖

    Abraham Silberschatz, Peter B. Galvin, and Gerg Gagne, (2014) ―Operating System Concepts ‖. Willey India 9th Edition, pp 944

    work page 2014

  2. [2]

    Tanenbaum, and Herbert Bos, (2014) ―Modern Operating System‖

    Andrew S. Tanenbaum, and Herbert Bos, (2014) ―Modern Operating System‖. Pearson 4th Edition, pp 1080

    work page 2014

  3. [3]

    Dhamdhere, (2006) ―Operating Systems: A Concept -based Approach‖

    Dhananjay M. Dhamdhere, (2006) ―Operating Systems: A Concept -based Approach‖. Mc Graw Hill Third Edition

    work page 2006

  4. [4]

    International Journal of Computer Science Engineering and Information Technology (IJCSEIT), Vol

    Amar Ranjan Dash, Sandipta Kumar Sahu, and Sanjay Kumar Samantra, (2015) ―An optimized round Robin CPU scheduling algorithm with dynamic time quantum‖. International Journal of Computer Science Engineering and Information Technology (IJCSEIT), Vol . 5(1), AIRCC, PP 7 -26, DOI: 10.5121/ijcseit.2015.5102, arXiv preprint arXiv:1605.00362. International Journ...

    work page arXiv 2015

  5. [5]

    International Journal

    Amar Ranjan Dash, Sandipta Kumar Sahu, Sanjay Kumar Samantra, and Sradhanjali Sabat, (2015) ―Characteristic specific prioritized dynamic average burst round robi n scheduling for uniprocessor and multiprocessor environment‖. International Journal. Of Computer Science Engineering and Applications (IJCSEA), Vol 5(5), AIRCC, PP 1 -20, DOI: 10.5121/ijcsea.201...

    work page arXiv 2015

  6. [6]

    Performance Analysis of Disk Scheduling Algorithms

    C. Mallikarjuna, and P. Chitti Babu, (2016) "Performance Analysis of Disk Scheduling Algorithms", International Journal of Computer Sciences and Engineering, Vol. 4(5), pp 180 -184

    work page 2016

  7. [7]

    A Disk Scheduling Algorithm Simulator

    Sukanya Suranauwarat, (2007) "A Disk Scheduling Algorithm Simulator." The AS EE Computers in Education (CoED) Journal, Vol. 8(3), pp 1-9

    work page 2007

  8. [8]

    International Journal of Computer Applications, Vol

    Manish Kumar Mishra, (2012) ―An improved FCFS (IFCFS) disk scheduling algorithm‖. International Journal of Computer Applications, Vol. 47(13), pp 20 -24

    work page 2012

Show all 21 references
  1. [9]

    In Proceedings of the winter 1990 USENIX technical conference, pp

    Margo Seltzer, Peter Chen, and John Ousterhout, (1990, January) ―Disk scheduling revisited‖. In Proceedings of the winter 1990 USENIX technical conference, pp. 313 -323

    1990

  2. [10]

    International Journal of Computer Applications, Vol

    Saman Rasool, and Ritika Gakher, (2015) ―Reformed FCFS Disk Scheduling Algorithm‖. International Journal of Computer Applications, Vol. 127(13), pp 38-41

    2015

  3. [11]

    International Journal of Computer Applications, pp 28-35, arXiv preprint arXiv:1403.0334

    Sourav Kumar Bhoi, Sanjany Kumar Panda, and Imran Hossain Faruk, (2012) ―Design and Performance Evaluation of an Optimized Disk Scheduling Algorithm (ODSA)‖. International Journal of Computer Applications, pp 28-35, arXiv preprint arXiv:1403.0334

    2012 arXiv

  4. [12]

    Nasim Akhter, and Mohammod Abdul Kashem, (2013) ―A New Heuristic Disk Scheduling Algorithm‖

    Sandipon Saha, Md. Nasim Akhter, and Mohammod Abdul Kashem, (2013) ―A New Heuristic Disk Scheduling Algorithm‖. International Journal of Scientific & Technology and Research, Vol. 2(1), 2013, pp 49-53

    2013

  5. [13]

    M. R. Mahesh Kumar, B. Renuka Rajendra, (2015) ―An improved approach to maximize the performance of disk scheduling algorithm by minimizing the head movement and seek time using sort mid current comparison (SMCC) algorithm‖. In Proceedings of 3rd Inter national Conference on R...

    2015 doi

  6. [14]

    International Journal of Computer Application, Vol

    Jainil Vachhani and Yash Turakhia, (2017) ―Design and Performance Evaluation of Median Range Scheduling Algorithm‖. International Journal of Computer Application, Vol. 172 (4), pp 6 -8

    2017

  7. [15]

    Journal of Computer and Communications, vol

    Priya Hooda, and Supriya Raheja, (2014) ―A new approach to disk scheduling using fuzzy logic‖. Journal of Computer and Communications, vol. 2, pp 1-5, DOI: 10.4236/jcc.2014.21001

    2014

  8. [16]

    In Database Systems for Advanced Applications, pp 395 -402, DOI: 10.1142/9789814503730_0045

    Kitae Hwang and Heonshik Shin, (1993) ―New Disk Scheduling Algorithms for Reduced Rotational Latency‖. In Database Systems for Advanced Applications, pp 395 -402, DOI: 10.1142/9789814503730_0045

    1993 doi

  9. [17]

    In Proceeding Fifth International Conference on Advanced Computing & Communication Technologies (ACCT -2015), IEEE, pp

    Karishma Singh, NIdhi, Divya Rastogi, and Day ashankar Singh, (2015) ―Optimized Two Head Disk Scheduling Algorithm (OTHDSA)‖. In Proceeding Fifth International Conference on Advanced Computing & Communication Technologies (ACCT -2015), IEEE, pp. 234 -240, DOI: 10.1109/ACCT.2015.70

    2015 doi

  10. [18]

    Comparative Study of Disk Scheduling Algorithms and Proposal of a New Algorithm for Better Efficiency

    Avneesh Shankar, Abhijeet Ravat, and Abhishek Kumar Pandey. (2019) "Comparative Study of Disk Scheduling Algorithms and Proposal of a New Algorithm for Better Efficiency". In Proceedings of 2nd International Conference on Advanced Computing and Software Engineering (ICA CSE) 2...

    2019 doi

  11. [19]

    ACM SIGARCH Computer Architecture News, Vol

    Alexander Thomasian, ―Survey and analysis of disk scheduling methods‖. ACM SIGARCH Computer Architecture News, Vol. 39(2), 2011, pp 8-25. International Journal of Computer Science, Engineering and Applications (IJCSEA) Vol. 9, No. 1/2/3, June 2019 21

    2011

  12. [20]

    In 2008 IEEE International Symposium on Modeling, Analysis and Simulation of Computers and Telecommunication Systems, IEEE, (2008, September) p p

    Anthony Hylick, Ripduman Sohan, Andrew Rice, and Brian Jones, ―An analysis of hard drive energy consumption‖. In 2008 IEEE International Symposium on Modeling, Analysis and Simulation of Computers and Telecommunication Systems, IEEE, (2008, September) p p. 1-10

    2008

  13. [21]

    coursework for PH240 Stanford University, 2018, http://large.stanford.edu/courses/2018/ph240/jiang2/ (accessed on 09/06/2019) Authors Amar Ranjan Dash obtained his B

    Wentao Jiang, ―Energy to Store One Bit‖. coursework for PH240 Stanford University, 2018, http://large.stanford.edu/courses/2018/ph240/jiang2/ (accessed on 09/06/2019) Authors Amar Ranjan Dash obtained his B. Tech Comp. Sc. from Biju P atnaik University of Technology, India and...

    2018

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.