pith. sign in

arxiv: 2404.00507 · v3 · pith:Q3D357U6new · submitted 2024-03-31 · 💻 cs.OS · cs.DC

THEMIS: Time, Heterogeneity, and Energy Minded Scheduling for Fair Multi-Tenant Use in FPGAs

Emre Karabulut , Arsalan Ali Malik , Amro Awad , Aydin Aysu This is my paper

Pith reviewed 2026-05-24 02:43 UTC · model grok-4.3

classification 💻 cs.OS cs.DC
keywords FPGA multi-tenancyfair schedulingspatiotemporal fairnessenergy efficiencypartial reconfigurationheterogeneous regionscloud resource management
0
0 comments X

The pith

THEMIS scheduling algorithm achieves 24.2 to 98.4 percent better fairness in multi-tenant FPGAs by fixing time, energy, and heterogeneity issues.

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

The paper claims that earlier methods for fair scheduling on shared FPGAs relied on incorrect metrics and assumptions that ignored real hardware limits. THEMIS fixes this by tracking both when and where tasks run, factoring in energy use when deciding how long to run each tenant, and handling the fact that FPGA regions can be combined or split in limited ways. If true, this would let cloud systems share expensive FPGA hardware more evenly without one user hogging resources or wasting power. A reader should care because FPGAs are increasingly used in data centers for acceleration, and unfair sharing could limit their adoption. The single-board tests show clear gains over older approaches.

Core claim

The authors present THEMIS as a scheduling solution that ensures spatiotemporal fairness by considering both spatial and temporal resource allocation instead of assuming uniform task latency, incorporates energy by adjusting intervals and overheads, and accounts for heterogeneous regions and partial reconfiguration constraints on merging and splitting. On the Xilinx Zedboard, this yields fairness improvements of 24.2 to 98.4 percent compared to prior algorithms while enabling trade-offs such as 55.3 times energy versus 69.3 times fairness.

What carries the argument

THEMIS algorithm that integrates spatiotemporal fairness, energy-aware interval adjustment, and constraints from heterogeneous partial reconfiguration regions.

If this is right

  • Prior scheduling methods are outperformed in fairness by 24.2 to 98.4 percent.
  • Energy efficiency can be traded against fairness at factors of 55.3x versus 69.3x.
  • The approach respects real FPGA limits on region merging and splitting that earlier work overlooked.
  • Cloud providers gain a method to balance fairness and energy in multi-tenant FPGA setups.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • Extending the energy and heterogeneity logic to other accelerators like GPUs might yield similar fairness gains.
  • Production workloads with varying task latencies could amplify or reduce the reported improvements.
  • Scaling the evaluation to multi-board or larger FPGA systems would test if the trade-offs remain stable.
  • The method could influence how partial reconfiguration is managed in future FPGA cloud services.

Load-bearing premise

Results from testing on one specific Xilinx Zedboard board with its reconfiguration rules and energy model apply to other FPGAs and actual cloud workloads.

What would settle it

Running the same workloads on a different FPGA platform or with real multi-tenant traces that have non-uniform latencies and seeing if fairness gains disappear would disprove the generalization.

Figures

Figures reproduced from arXiv: 2404.00507 by Amro Awad, Arsalan Ali Malik, Aydin Aysu, Emre Karabulut.

Figure 1
Figure 1. Figure 1: Fairness and energy efficiency comparison of our work with prior [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: The proposed work’s desired average allocation (red dashed line) and [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: An example illustrating the slot allocation using the proposed [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Average slot allocation of proposed work on three FPGA slots with allocation interval set as [PITH_FULL_IMAGE:figures/full_fig_p007_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: The scheduling algorithms’ results for (a) FPGA slot utilization and [PITH_FULL_IMAGE:figures/full_fig_p008_5.png] view at source ↗
Figure 7
Figure 7. Figure 7: A comparative analysis of our proposed algorithm with prior work, [PITH_FULL_IMAGE:figures/full_fig_p009_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: The proposed work was able to maintain fairness even in a scenario [PITH_FULL_IMAGE:figures/full_fig_p010_8.png] view at source ↗
read the original abstract

Using correct design metrics and understanding the limitations of the underlying technology is critical to developing effective scheduling algorithms. Unfortunately, existing scheduling techniques used \emph{incorrect} metrics and had \emph{unrealistic} assumptions for fair scheduling of multi-tenant FPGAs where each tenant is aimed to share approximately the same number of resources both spatially and temporally. This paper introduces an enhanced fair scheduling algorithm for multi-tenant FPGA use, addressing previous metric and assumption issues, with three specific improvements claimed First, our method ensures spatiotemporal fairness by considering both spatial and temporal aspects, addressing the limitation of prior work that assumed uniform task latency. Second, we incorporate energy considerations into fairness by adjusting scheduling intervals and accounting for energy overhead, thereby balancing energy efficiency with fairness. Third, we acknowledge overlooked aspects of FPGA multi-tenancy, including heterogeneous regions and the constraints on dynamically merging/splitting partially reconfigurable regions. We develop and evaluate our improved fair scheduling algorithm with these three enhancements. Inspired by the Greek goddess of law and personification of justice, we name our fair scheduling solution THEMIS: \underline{T}ime, \underline{H}eterogeneity, and \underline{E}nergy \underline{Mi}nded \underline{S}cheduling. We used the Xilinx Zedboard XC7Z020 to quantify our approach's savings. Compared to previous algorithms, our improved scheduling algorithm enhances fairness between 24.2--98.4\% and allows a trade-off between 55.3$\times$ in energy vs. 69.3$\times$ in fairness. The paper thus informs cloud providers about future scheduling optimizations for fairness with related challenges and opportunities.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 2 minor

Summary. The paper introduces THEMIS, an enhanced fair scheduling algorithm for multi-tenant FPGAs that incorporates spatiotemporal fairness (addressing non-uniform task latencies), energy overheads in scheduling intervals, and constraints from heterogeneous partially reconfigurable regions and dynamic merging/splitting. It evaluates the approach on a Xilinx Zedboard XC7Z020, claiming 24.2--98.4% fairness improvements over prior algorithms and energy-fairness trade-offs of 55.3× versus 69.3×.

Significance. If the measured gains prove robust, the work could usefully inform cloud FPGA schedulers by highlighting the interplay of energy, heterogeneity, and PR-region constraints. The explicit treatment of non-uniform latencies and region merging is a concrete advance over prior metric choices.

major comments (2)
  1. [§5] §5 (Evaluation) and abstract: the headline claims of 24.2--98.4% fairness improvement and 55.3×/69.3× trade-offs rest exclusively on measurements from the single Xilinx Zedboard XC7Z020 under its specific PR-region sizes, reconfiguration latencies, and energy model; no sensitivity study or results on larger platforms (Virtex Ultrascale+, AWS F1) are supplied, which directly undermines the generalization of the quantitative results.
  2. [§4] §4 (Algorithm description): no pseudocode, explicit baseline definitions, or closed-form fairness metric is provided, making it impossible to verify how the three claimed enhancements (spatiotemporal fairness, energy adjustment, heterogeneity handling) produce the reported deltas; this is load-bearing for the central contribution.
minor comments (2)
  1. [Abstract] Abstract and §1: the phrase 'enhances fairness between 24.2--98.4%' is ambiguous; clarify whether this is relative improvement, absolute delta, or range across workloads.
  2. [§5] Figure captions and §5: workload details (task mix, arrival rates, region sizes) are referenced but not tabulated; add a summary table for reproducibility.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive comments, which help clarify the presentation of our contributions. We address each major point below and indicate the planned revisions.

read point-by-point responses

  1. Referee: [§5] §5 (Evaluation) and abstract: the headline claims of 24.2--98.4% fairness improvement and 55.3×/69.3× trade-offs rest exclusively on measurements from the single Xilinx Zedboard XC7Z020 under its specific PR-region sizes, reconfiguration latencies, and energy model; no sensitivity study or results on larger platforms (Virtex Ultrascale+, AWS F1) are supplied, which directly undermines the generalization of the quantitative results.

    Authors: We agree that all quantitative results are derived from the Xilinx Zedboard XC7Z020. This platform was chosen because it is a widely available device supporting partial reconfiguration with heterogeneous regions, allowing direct measurement of the spatiotemporal and energy effects under realistic constraints. The relative improvements (24.2--98.4%) arise from the new metrics rather than platform-specific tuning. In the revised manuscript we will add a dedicated subsection in §5 discussing how the fairness metric, energy-interval adjustment, and region-merging logic scale to larger devices such as Virtex Ultrascale+ and AWS F1, including qualitative analysis of reconfiguration latency and region-size effects. New experimental results on those platforms are not feasible within the current hardware access, so the quantitative claims will remain tied to the evaluated board while the algorithmic contributions are framed more generally. revision: partial

  2. Referee: [§4] §4 (Algorithm description): no pseudocode, explicit baseline definitions, or closed-form fairness metric is provided, making it impossible to verify how the three claimed enhancements (spatiotemporal fairness, energy adjustment, heterogeneity handling) produce the reported deltas; this is load-bearing for the central contribution.

    Authors: We accept this criticism. The current manuscript describes the three enhancements in prose but omits pseudocode, precise baseline definitions, and the closed-form expression for the spatiotemporal fairness metric. In the revised version we will insert (i) a new Algorithm 1 box containing the full THEMIS scheduling procedure, (ii) explicit definitions of the three prior algorithms used for comparison, and (iii) the mathematical formulation of the fairness metric that incorporates both spatial occupancy and temporal latency variation. These additions will make the source of the reported deltas verifiable without altering the experimental results. revision: yes

Circularity Check

0 steps flagged

No circularity; claims are empirical hardware measurements, not self-referential definitions or fitted predictions

full rationale

The paper's central claims (fairness improvements of 24.2–98.4% and energy/fairness trade-offs) are presented as direct outcomes of running the THEMIS scheduler on a Xilinx Zedboard XC7Z020 under measured partial-reconfiguration and energy conditions. No equations define a metric in terms of the scheduler output and then treat the output as an independent prediction; no self-citation chain is invoked to justify uniqueness or force the result; the algorithm improvements are described as explicit handling of spatial/temporal/energy/heterogeneity factors rather than ansatzes smuggled from prior author work. The derivation chain is therefore self-contained against external benchmarks (the physical device runs).

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

No free parameters or invented entities are described; the work relies on standard domain assumptions about FPGA partial reconfiguration.

axioms (1)
  • domain assumption FPGA partial reconfiguration regions have fixed heterogeneity and merging/splitting constraints that must be respected by any scheduler.
    Invoked when the paper states it acknowledges overlooked aspects of FPGA multi-tenancy including heterogeneous regions and constraints on dynamically merging/splitting regions.

pith-pipeline@v0.9.0 · 5856 in / 1221 out tokens · 26394 ms · 2026-05-24T02:43:40.263264+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

42 extracted references · 42 canonical work pages · 2 internal anchors

  1. [1]

    Sharing, Protection, and Compatibility for Reconfigurable Fabric with A MORPHOS,

    A. Khawaja, J. Landgraf, R. Prakash, M. Wei, E. Schkufza, and C. J. Rossbach, “Sharing, Protection, and Compatibility for Reconfigurable Fabric with A MORPHOS,” in 13th USENIX Symposium on Operating Systems Design and Implementation (OSDI 18) , 2018, pp. 107–127

    work page 2018

  2. [2]

    Do OS abstractions make sense on FPGAs?

    D. Korolija, T. Roscoe, and G. Alonso, “Do OS abstractions make sense on FPGAs?” in 14th USENIX Symposium on Operating Systems Design and Implementation (OSDI 20) , 2020, pp. 991–1010

    work page 2020

  3. [3]

    Virtualized FPGA Acceler- ators for Efficient Cloud Computing,

    S. A. Fahmy, K. Vipin, and S. Shreejith, “Virtualized FPGA Acceler- ators for Efficient Cloud Computing,” in 2015 IEEE 7th International Conference on Cloud Computing Technology and Science (CloudCom) . IEEE, 2015, pp. 430–435

    work page 2015

  4. [4]

    Virtualizing FPGAs in the Cloud,

    Y . Zha and J. Li, “Virtualizing FPGAs in the Cloud,” in ACM Interna- tional Conference on Architectural Support for Programming Languages and Operating Systems (ASPLOS) , 2020, pp. 845–858

    work page 2020

  5. [5]

    A Hypervisor for Shared-Memory FPGA Platforms,

    J. Ma, G. Zuo, K. Loughlin, X. Cheng, Y . Liu, A. M. Eneyew, Z. Qi, and B. Kasikci, “A Hypervisor for Shared-Memory FPGA Platforms,” in Proceedings of the Twenty-Fifth International Conference on Archi- tectural Support for Programming Languages and Operating Systems , 2020, pp. 827–844

    work page 2020

  6. [6]

    Resource Elastic Virtualization for FPGAs using OpenCL,

    A. Vaishnav, K. D. Pham, D. Koch, and J. Garside, “Resource Elastic Virtualization for FPGAs using OpenCL,” in IEEE International Con- ference on Field Programmable Logic and Applications (FPL) , 2018, pp. 111–1117

    work page 2018

  7. [7]

    Looking beyond GPUs for DNN Scheduling on Multi-Tenant Clusters,

    J. Mohan, A. Phanishayee, J. Kulkarni, and V . Chidambaram, “Looking beyond GPUs for DNN Scheduling on Multi-Tenant Clusters,” in 16th USENIX Symposium on Operating Systems Design and Implementation (OSDI 22), 2022, pp. 579–596

    work page 2022

  8. [8]

    A Survey of Multi-tenant Deep Learning Inference on GPU,

    F. Yu, D. Wang, L. Shangguan, M. Zhang, C. Liu, and X. Chen, “A Survey of Multi-tenant Deep Learning Inference on GPU,” arXiv preprint arXiv:2203.09040, 2022

    work page arXiv 2022

  9. [9]

    Performance Isolation and Fairness for Multi-tenant Cloud Storage,

    D. Shue, M. J. Freedman, and A. Shaikh, “Performance Isolation and Fairness for Multi-tenant Cloud Storage,” inpresented as part of the 10th USENIX Symposium on Operating Systems Design and Implementation (OSDI 12), 2012, pp. 349–362

    work page 2012

  10. [10]

    SQLVM: Performance Isolation in Multi-Tenant Relational Database- as-a-Service,

    V . Narasayya, S. Das, M. Syamala, B. Chandramouli, and S. Chaudhuri, “SQLVM: Performance Isolation in Multi-Tenant Relational Database- as-a-Service,” in CIDR 2013, 2013

    work page 2013

  11. [11]

    QoS Aware Design-Time/Run-Time Manager for FPGA-Based Embedded Systems,

    A. Duhamel and S. Pillement, “QoS Aware Design-Time/Run-Time Manager for FPGA-Based Embedded Systems,” in Design and Archi- tecture for Signal and Image Processing: 15th International Workshop, DASIP 2022, Budapest, Hungary, June 20–22, 2022, Proceedings . Springer, 2022, pp. 96–107

    work page 2022

  12. [12]

    Benefits of Partial Reconfiguration,

    C. Kao, “Benefits of Partial Reconfiguration,” Xcell journal, vol. 55, pp. 65–67, 2005. 5https://github.com/aamalik3/THEMIS.git

    work page 2005

  13. [13]

    Koch, Partial reconfiguration on FPGAs: Architectures, Tools and Applications

    D. Koch, Partial reconfiguration on FPGAs: Architectures, Tools and Applications. Springer Science & Business Media, 2012, vol. 153

    work page 2012

  14. [14]

    Spatiotemporal Strategies for Long-Term FPGA Resource Management,

    A. Mehrabi, D. J. Sorin, and B. C. Lee, “Spatiotemporal Strategies for Long-Term FPGA Resource Management,” in 2022 IEEE International Symposium on Performance Analysis of Systems and Software (ISPASS). IEEE, 2022, pp. 198–209

    work page 2022

  15. [15]

    Programming and Runtime Support to Blaze FPGA Accelerator Deployment at Datacenter Scale,

    M. Huang, D. Wu, C. H. Yu, Z. Fang, M. Interlandi, T. Condie, and J. Cong, “Programming and Runtime Support to Blaze FPGA Accelerator Deployment at Datacenter Scale,” in Proceedings of the Seventh ACM Symposium on Cloud Computing , 2016, pp. 456–469

    work page 2016

  16. [16]

    RC3E: Provision and Management of Reconfigurable Hardware Accelerators in a Cloud Environment

    O. Knodel and R. G. Spallek, “RC3E: Provision and Management of Reconfigurable Hardware Accelerators in a Cloud Environment v,”arXiv preprint arXiv:1508.06843, 2015

    work page internal anchor Pith review Pith/arXiv arXiv 2015

  17. [17]

    UG947 Vivado Design Suite Tutorial Dynamic Function eXchange , Xilinx Inc, 4 2022, v2021.2

    work page 2022

  18. [18]

    EPOCH: Enabling Preemption Operation for Context Saving in Heterogeneous FPGA Systems

    Malik, Arsalan Ali and Karabulut, Emre and Aysu, Aydin, “EPOCH: Enabling Preemption Operation for Context Saving in Heterogeneous FPGA Systems,” arXiv preprint arXiv:2501.16205 , 2025

    work page internal anchor Pith review Pith/arXiv arXiv 2025

  19. [19]

    PR Crisis: Analyzing and Fixing Partial Reconfiguration in Multi-Tenant Cloud FPGAs,

    E. Karabulut, C. Yuvarajappa, M. I. Shaik, S. Potluri, A. Awad, and A. Aysu, “PR Crisis: Analyzing and Fixing Partial Reconfiguration in Multi-Tenant Cloud FPGAs,” in ACM Workshop on Attacks and Solutions in Hardware Security (ASHES) , 2022, pp. 101–106

    work page 2022

  20. [20]

    Dynamic and Partial Reconfiguration Power Con- sumption Runtime Measurements Analysis for ZYNQ SoC Devices ,

    M. A. Rihani et al, “Dynamic and Partial Reconfiguration Power Con- sumption Runtime Measurements Analysis for ZYNQ SoC Devices ,” in IEEE International Symposium on Wireless Communication Systems , 2016, pp. 592–596

    work page 2016

  21. [21]

    Cuttlesys: Data-driven resource management for interactive services on Reconfigurable Multicores,

    N. Kulkarni, G. Gonzalez-Pumariega, A. Khurana, C. A. Shoemaker, C. Delimitrou, and D. H. Albonesi, “Cuttlesys: Data-driven resource management for interactive services on Reconfigurable Multicores,” in 2020 53rd Annual IEEE/ACM International Symposium on Microarchi- tecture (MICRO). IEEE, 2020, pp. 650–664

    work page 2020

  22. [22]

    REF: Resource Elasticity Fairness with Sharing Incentives for Multiprocessors,

    S. M. Zahedi and B. C. Lee, “REF: Resource Elasticity Fairness with Sharing Incentives for Multiprocessors,” ACM Sigplan Notices , vol. 49, no. 4, pp. 145–160, 2014

    work page 2014

  23. [23]

    Dominant Resource Fairness: Fair Allocation of Multiple Resource Types,

    A. Ghodsi, M. Zaharia, B. Hindman, A. Konwinski, S. Shenker, and I. Stoica, “Dominant Resource Fairness: Fair Allocation of Multiple Resource Types,” in 8th USENIX symposium on networked systems design and implementation (NSDI 11) , 2011

    work page 2011

  24. [24]

    Beyond Dominant Resource Fairness: Extensions, Limitations, and Indivisibilities,

    D. C. Parkes, A. D. Procaccia, and N. Shah, “Beyond Dominant Resource Fairness: Extensions, Limitations, and Indivisibilities,” ACM Transactions on Economics and Computation (TEAC) , vol. 3, no. 1, pp. 1–22, 2015

    work page 2015

  25. [25]

    Round-Robin Scheduling for Max-Min Fairness in Data Network,

    E. L. Hahne, “Round-Robin Scheduling for Max-Min Fairness in Data Network,” IEEE Journal on Selected Areas in communications , vol. 9, no. 7, pp. 1024–1039, 1991

    work page 1991

  26. [26]

    Schedule-Induced Polydipsia as a Function of Fixed Interval Length,

    J. L. Falk, “Schedule-Induced Polydipsia as a Function of Fixed Interval Length,” Journal of the experimental analysis of Behavior , vol. 9, no. 1, pp. 37–39, 1966

    work page 1966

  27. [27]

    MachSuite: Benchmarks for Accelerator Design and Customized Architectures,

    B. Reagen, R. Adolf, Y . S. Shao, G.-Y . Wei, and D. Brooks, “MachSuite: Benchmarks for Accelerator Design and Customized Architectures,” in 2014 IEEE International Symposium on Workload Characterization (IISWC). IEEE, 2014, pp. 110–119

    work page 2014

  28. [28]

    ZyCAP: Efficient Partial Reconfiguration Management on the Xilinx Zynq,

    K. Vipin and S. A. Fahmy, “ZyCAP: Efficient Partial Reconfiguration Management on the Xilinx Zynq,” IEEE Embedded Systems Letters , vol. 6, no. 3, pp. 41–44, 2014

    work page 2014

  29. [29]

    A Novel Mechanism for Effective Hardware Task Preemption in Dynamically Reconfigurable Systems,

    K. Jozwik, H. Tomiyama, S. Honda, and H. Takada, “A Novel Mechanism for Effective Hardware Task Preemption in Dynamically Reconfigurable Systems,” in 2010 International Conference on Field Programmable Logic and Applications . IEEE, 2010, pp. 352–355

    work page 2010

  30. [30]

    VR-ZYCAP: A Versatile Resource-Level ICAP Controller for ZYNQ SOC,

    B. Sultana, A. Ullah, A. A. Malik, A. Zahir, P. Reviriego, F. B. Muslim, N. Ullah, and W. Ahmad, “VR-ZYCAP: A Versatile Resource-Level ICAP Controller for ZYNQ SOC,” Electronics, vol. 10, no. 8, p. 899, 2021

    work page 2021

  31. [31]

    Run-time Partial Reconfigura- tion Speed Investigation and Architectural Design Space Exploration,

    M. Liu, W. Kuehn, Z. Lu, and A. Jantsch, “Run-time Partial Reconfigura- tion Speed Investigation and Architectural Design Space Exploration,” in 2009 International Conference on Field Programmable Logic and Applications. IEEE, 2009, pp. 498–502

    work page 2009

  32. [32]

    Runtime FPGA Partial Reconfiguration,

    E. J. McDonald, “Runtime FPGA Partial Reconfiguration,” in 2008 IEEE Aerospace Conference. IEEE, 2008, pp. 1–7

    work page 2008

  33. [33]

    A High Speed Open Source Controller for FPGA Partial Reconfiguration,

    K. Vipin and S. A. Fahmy, “A High Speed Open Source Controller for FPGA Partial Reconfiguration,” in 2012 International Conference on Field-Programmable Technology. IEEE, 2012, pp. 61–66

    work page 2012

  34. [34]

    FaRM: Fast Reconfiguration Manager for Reducing Reconfiguration Time Overhead on FPGA,

    F. Duhem, F. Muller, and P. Lorenzini, “FaRM: Fast Reconfiguration Manager for Reducing Reconfiguration Time Overhead on FPGA,” in Reconfigurable Computing: Architectures, Tools and Applications: 7th International Symposium, ARC 2011, Belfast, UK, March 23-25, 2011. Proceedings 7. Springer, 2011, pp. 253–260

    work page 2011

  35. [35]

    Achieving Energy Efficiency through Runtime Partial Reconfiguration on Reconfigurable Systems,

    S. Liu, R. N. Pittman, A. Forin, and J.-L. Gaudiot, “Achieving Energy Efficiency through Runtime Partial Reconfiguration on Reconfigurable Systems,” ACM Transactions on Embedded Computing Systems (TECS), vol. 12, no. 3, pp. 1–21, 2013

    work page 2013

  36. [36]

    ZyPR: End-to-End Build Tool and Runtime Manager for Partial Reconfiguration of FPGA SoCs at the Edge,

    A. R. Bucknall and S. A. Fahmy, “ZyPR: End-to-End Build Tool and Runtime Manager for Partial Reconfiguration of FPGA SoCs at the Edge,” ACM Transactions on Reconfigurable Technology and Systems , 2023

    work page 2023

  37. [37]

    Enabling secure and efficient sharing of accelerators in expeditionary systems,

    A. A. Malik, E. Karabulut, A. Awad, and A. Aysu, “Enabling secure and efficient sharing of accelerators in expeditionary systems,” Journal of Hardware and Systems Security , vol. 8, no. 2, pp. 94–112, 2024

    work page 2024

  38. [38]

    Isolation design flow effectiveness evaluation methodology for Zynq SoCs,

    Malik, Arsalan Ali and Ullah, Anees and Zahir, Ali and others, “Isolation design flow effectiveness evaluation methodology for Zynq SoCs,” Electronics, 2020

    work page 2020

  39. [39]

    Amazon Web Services,

    Amazon, EC, “Amazon Web Services,” Available in: http://aws. amazon. com/es/ec2/(November 2012), 2015

    work page 2012

  40. [40]

    Springer, 2015

    Copeland, Marshall and Soh, Julian and Puca, Anthony and Manning, Mike and Gollob, David and Copeland, Marshall and Soh, Julian and Puca, Anthony and Manning, Mike and Gollob, David, Microsoft Azure and Cloud Computing . Springer, 2015

    work page 2015

  41. [41]

    A review on amazon web service (aws), microsoft azure & google cloud platform (gcp) services,

    Gupta, Bulbul and Mittal, Pooja and Mufti, Tabish, “A review on amazon web service (aws), microsoft azure & google cloud platform (gcp) services,” in Proceedings of the 2nd International Conference on ICT for Digital, Smart, and Sustainable Development, ICIDSSD 2020, 27-28 February 2020, Jamia Hamdard, New Delhi, India , 2021

    work page 2020

  42. [42]

    SoK: Secure FPGA multi-tenancy in the cloud: Challenges and opportunities,

    Dessouky, Ghada and Sadeghi, Ahmad-Reza and Zeitouni, Shaza, “SoK: Secure FPGA multi-tenancy in the cloud: Challenges and opportunities,” in 2021 IEEE European Symposium on Security and Privacy (EuroS&P). IEEE, 2021

    work page 2021