pith. sign in

arxiv: 2505.21899 · v3 · submitted 2025-05-28 · 💻 cs.DC

Jointλ: Orchestrating Serverless Workflows on Jointcloud FaaS Systems

Rui Li , Jianfei Liu , Zhilin Yang , Peichang Shi , Guodong Yi , Huaimin Wang This is my paper

Pith reviewed 2026-05-19 14:18 UTC · model grok-4.3

classification 💻 cs.DC
keywords serverless workflowsmulti-cloud orchestrationFaaS systemsdistributed runtimefault tolerancemakespan reductioncost optimization
0
0 comments X

The pith

Jointλ runs serverless workflows across separate cloud platforms by moving orchestration into the functions themselves.

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

The paper sets out to prove that serverless workflows can be coordinated across different cloud providers without a single central controller. It does this by adding a thin compatibility layer called Backend-Shim and by letting each function handle its own invocations and data moves. A sympathetic reader would care because the approach removes vendor lock-in while cutting execution time and cost at the same time. The system also adds datastore-based mechanisms that keep execution reliable even when individual clouds or links fail. Tests on AWS and Aliyun with four workflows supply the concrete evidence for these gains.

Core claim

Jointλ is a distributed runtime that orchestrates serverless workflows on Jointcloud FaaS systems. It introduces the Backend-Shim compatibility layer to exploit differences between clouds for better scheduling and lower bills under on-demand pricing. By replacing centralized nodes with function-side orchestration, each function can invoke successors and move data independently, which removes most cross-cloud coordination traffic. Exactly-once semantics are maintained through persistent datastores and failover paths that survive partial cloud outages.

What carries the argument

The Backend-Shim compatibility layer together with function-side orchestration, which together replace a central coordinator with local decisions at each function.

If this is right

  • Workflows finish up to 3.3 times sooner than leading single-cloud commercial services.
  • Cost drops by as much as 65 percent by routing functions to whichever cloud is cheaper at each step.
  • Execution stays correct even after cloud or link failures because of the exactly-once datastore guarantees.
  • The same system runs up to 4 times faster than prior cross-cloud orchestrators while keeping competitive cost.

Where Pith is reading between the lines

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

  • The same local-orchestration pattern could be applied to other multi-provider settings such as edge-cloud combinations.
  • Application developers might begin writing workflows that explicitly encode price or latency preferences so the shim layer can exploit them automatically.
  • Future serverless platforms could adopt similar shim layers as a standard way to support migration between providers without rewriting code.

Load-bearing premise

The four tested workflows and the two chosen FaaS platforms are representative of wider cross-cloud differences in performance, pricing, and network behavior.

What would settle it

Measure the same four workflows on three additional cloud providers under deliberately varied network latency and bandwidth; if the reported speedups and cost savings shrink below 1.5 times, the central performance claim would not hold.

Figures

Figures reproduced from arXiv: 2505.21899 by Guodong Yi, Huaimin Wang, Jianfei Liu, Peichang Shi, Rui Li, Zhilin Yang.

Figure 16
Figure 16. Figure 16: We conservatively use the higher price between Amazon DynamoDB and Aliyun TableStore, i.e., 1.4269$ per 1M [PITH_FULL_IMAGE:figures/full_fig_p018_16.png] view at source ↗
read the original abstract

Existing serverless workflow orchestration systems are predominantly designed for a single-cloud FaaS system, leading to vendor lock-in. This restricts performance optimization, cost reduction, and availability of applications. However, orchestrating serverless workflows on Jointcloud FaaS systems faces two main challenges: (1) additional overhead caused by centralized cross-cloud orchestration; and (2) a lack of reliable failover and fault-tolerant mechanisms for cross-cloud serverless workflows. To address these challenges, we propose Joint$\lambda$, a distributed runtime system designed to orchestrate serverless workflows on multiple FaaS systems without relying on a centralized orchestrator. Joint$\lambda$ introduces a compatibility layer, Backend-Shim, leveraging inter-cloud heterogeneity to optimize makespan and reduce costs with on-demand billing. By using function-side orchestration instead of centralized nodes, it enables independent function invocations and data transfers, reducing cross-cloud communication overhead. For high availability, it ensures exactly-once execution via datastores and failover mechanisms for serverless workflows on Jointcloud FaaS systems. We validate Joint$\lambda$ on two heterogeneous FaaS systems, AWS and Aliyun, with four workflows. Compared to the most advanced commercial orchestration services for single-cloud serverless workflows, Joint$\lambda$ reduces makespan by up to 3.3$\times$ while saving up to 65% in cost. Joint$\lambda$ is also up to 4.0$\times$ faster than state-of-the-art orchestrators for cross-cloud serverless workflows, while achieving competitive cost in representative scenarios and providing strong execution guarantees.

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 presents Jointλ, a distributed runtime system for orchestrating serverless workflows on Jointcloud FaaS systems spanning multiple clouds. It introduces a Backend-Shim compatibility layer to exploit inter-cloud heterogeneity, replaces centralized orchestration with function-side orchestration and independent data transfers, and provides exactly-once execution guarantees through datastore-based mechanisms and failover. Evaluation on AWS and Aliyun using four workflows reports up to 3.3× makespan reduction and 65% cost savings versus advanced single-cloud commercial services, plus up to 4.0× speedup versus state-of-the-art cross-cloud orchestrators.

Significance. If the reported gains hold, the work would meaningfully reduce vendor lock-in for serverless workflows by enabling practical multi-cloud execution with lower overhead and strong fault tolerance. Direct measurement on real platforms (AWS, Aliyun) and the parameter-free nature of the core architecture (no fitted constants in the orchestration logic) are strengths. However, the narrow evaluation base limits the assessed significance, as the headline claims rest on untested scaling assumptions.

major comments (2)
  1. Evaluation section: the 3.3× makespan and 4.0× speedup claims are measured exclusively on AWS+Aliyun with four workflows. No results are supplied for a third cloud, different latency distributions, or larger DAGs, leaving the general claim that cross-cloud transfer overhead remains sub-dominant unverified and load-bearing for the Jointcloud applicability stated in the abstract.
  2. Evaluation section: the reported speedups and cost numbers lack error bars, workload selection criteria, and a comparison against a properly tuned baseline that also performs direct cross-cloud transfers; these omissions prevent strong verification of the central performance claims.
minor comments (2)
  1. Abstract and §1: the phrase 'strong execution guarantees' should be defined more precisely (e.g., exactly-once semantics under what failure models) with a forward reference to the relevant section.
  2. Throughout: ensure consistent capitalization and definition of 'Backend-Shim' on first use and in all figures.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive and detailed review. The comments correctly identify areas where the evaluation can be strengthened to better support the claims. We address each major comment below and describe the changes planned for the revised manuscript.

read point-by-point responses

  1. Referee: Evaluation section: the 3.3× makespan and 4.0× speedup claims are measured exclusively on AWS+Aliyun with four workflows. No results are supplied for a third cloud, different latency distributions, or larger DAGs, leaving the general claim that cross-cloud transfer overhead remains sub-dominant unverified and load-bearing for the Jointcloud applicability stated in the abstract.

    Authors: We agree that the evaluation uses only two heterogeneous platforms (AWS and Aliyun) and four workflows. These were chosen because they exhibit substantial differences in latency, pricing, and invocation models, allowing us to demonstrate the benefits of Backend-Shim and function-side orchestration under realistic Jointcloud conditions. The architecture itself contains no cloud-specific constants and relies on independent data transfers, which we argue keeps transfer overhead sub-dominant even as the number of clouds grows. Nevertheless, we acknowledge that direct measurements on a third provider and larger DAGs would provide stronger verification. In the revision we will add results from a third cloud (Google Cloud Functions) and two additional larger workflows, together with a short scaling discussion. revision: yes

  2. Referee: Evaluation section: the reported speedups and cost numbers lack error bars, workload selection criteria, and a comparison against a properly tuned baseline that also performs direct cross-cloud transfers; these omissions prevent strong verification of the central performance claims.

    Authors: We thank the referee for this observation. The four workflows were selected as representative patterns drawn from prior serverless workflow literature (image processing, data analytics, and ML inference pipelines). We will add error bars computed from at least five independent runs per configuration. We will also expand the text to explicitly state the workload selection rationale and to clarify that the cross-cloud baseline orchestrators were configured to use direct inter-cloud transfers. If a more finely tuned variant of any baseline is needed, we will include it in the revised evaluation. revision: yes

Circularity Check

0 steps flagged

No circularity: performance claims rest on direct empirical measurements, not derivations or fitted models

full rationale

The paper proposes an architectural system (Backend-Shim, function-side orchestration, datastore-based exactly-once) and reports makespan/cost improvements from concrete runs on AWS and Aliyun using four specific workflows. No equations, first-principles derivations, or predictive models appear in the provided text; the 3.3×/4.0× figures are stated as observed outcomes of those runs rather than quantities computed from parameters fitted to the same data or justified solely by self-citation. The evaluation is therefore self-contained against external benchmarks and does not reduce any claimed result to its own inputs by construction.

Axiom & Free-Parameter Ledger

0 free parameters · 2 axioms · 1 invented entities

The central claim rests on the assumption that direct function-to-function data movement across clouds is feasible and that datastore-based exactly-once semantics can be implemented without introducing new bottlenecks; no free parameters or invented physical entities are introduced.

axioms (2)
  • domain assumption Heterogeneous FaaS platforms expose sufficient APIs for direct function invocation and data transfer without a central coordinator.
    Invoked when the paper claims function-side orchestration reduces cross-cloud overhead.
  • domain assumption Datastore writes are reliable enough to guarantee exactly-once execution across clouds.
    Required for the failover mechanism described in the abstract.
invented entities (1)
  • Backend-Shim no independent evidence
    purpose: Compatibility layer that hides differences between FaaS platforms to enable direct cross-cloud function calls and data movement.
    New software component introduced to support the distributed orchestration model.

pith-pipeline@v0.9.0 · 5837 in / 1523 out tokens · 36017 ms · 2026-05-19T14:18:57.817199+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

65 extracted references · 65 canonical work pages · 1 internal anchor

  1. [1]

    Cloud Programming Simplified: A Berkeley View on Serverless Computing.arXiv (Cornell University)

    Jonas E, Schleier-Smith J, Sreekanti V , et al. Cloud Programming Simplified: A Berkeley View on Serverless Computing.arXiv (Cornell University)

    work page

  2. [2]

    doi: 10.4230/oasics.microservices.2020-2022.5

    work page doi:10.4230/oasics.microservices.2020-2022.5 2020

  3. [3]

    Encoding, fast and slow:{Low-Latency} video processing using thousands of tiny threads

    Fouladi S, Wahby RS, Shacklett B, et al. Encoding, fast and slow:{Low-Latency} video processing using thousands of tiny threads. In: USENIX Association. 2017:363–376

    work page 2017

  4. [4]

    From laptop to lambda: outsourcing everyday jobs to thousands of transient functional containers

    Fouladi S, Romero F, Iter D, et al. From laptop to lambda: outsourcing everyday jobs to thousands of transient functional containers. In: USENIX. 2019:475–488

    work page 2019

  5. [5]

    Fault-tolerant and transactional stateful serverless workflows.Operating Systems Design and Implementation.2020:1187–1204

    Zhang H, Cardoza A, Chen PB, Angel S, Liu V . Fault-tolerant and transactional stateful serverless workflows.Operating Systems Design and Implementation.2020:1187–1204

    work page 2020

  6. [6]

    Doing more with less: Orchestrating serverless applications without an orchestrator

    Liu DH, Levy A, Noghabi S, Burckhardt S. Doing more with less: Orchestrating serverless applications without an orchestrator. In: USENIX Association. 2023:1505–1519

    work page 2023

  7. [7]

    FaaSFlow: enable efficient workflow execution for function-as-a-service

    Li Z, Liu Y , Guo L, et al. FaaSFlow: enable efficient workflow execution for function-as-a-service. In: ACM. 2022:782–796

    work page 2022

  8. [8]

    Durable functions: semantics for stateful serverless

    Burckhardt S, Gillum C, Justo D, Kallas K, McMahon C, Meiklejohn C. Durable functions: semantics for stateful serverless. In: . 5. ACM. 2021:1–27

    work page 2021

  9. [9]

    Sprocket

    Ao L, Izhikevich L, V oelker GM, Porter G. Sprocket. In: ACM. 2018:263–274 20 LiET AL

    work page 2018

  10. [10]

    Serverless execution of scientific workflows: Experiments with hyperflow, aws lambda and google cloud functions.Future Generation Computer Systems.2020;110:502–514

    Malawski M, Gajek A, Zima A, Balis B, Figiela K. Serverless execution of scientific workflows: Experiments with hyperflow, aws lambda and google cloud functions.Future Generation Computer Systems.2020;110:502–514

    work page 2020

  11. [11]

    Following the data, not the function: Rethinking function orchestration in serverless computing

    Yu M, Cao T, Wang W, Chen R. Following the data, not the function: Rethinking function orchestration in serverless computing. In: USENIX Association. 2023:1489–1504

    work page 2023

  12. [12]

    Characterizing commodity serverless computing platforms.Journal of Software: Evolution and Process

    Wen J, Liu Y , Chen Z, Chen J, Ma Y . Characterizing commodity serverless computing platforms.Journal of Software: Evolution and Process. 2023;35(10):2394–2417

    work page 2023

  13. [13]

    SeBS: A Serverless Benchmark Suite for Function-as-a-Service Computing.Zenodo (CERN European Organization for Nuclear Research).2021

    Copik M, Kwa´sniewski G, Besta M, Podstawski M, Hoefler T. SeBS: A Serverless Benchmark Suite for Function-as-a-Service Computing.Zenodo (CERN European Organization for Nuclear Research).2021. doi: 10.5281/zenodo.5209001

    work page doi:10.5281/zenodo.5209001 2021

  14. [14]

    Maissen P, Felber P, Kropf P, Schiavoni V . FaaSdom. In: ACM. 2020:73–84

    work page 2020

  15. [15]

    Characterizing serverless platforms with serverlessbench

    Yu T, Liu Q, Du D, et al. Characterizing serverless platforms with serverlessbench. In: ACM. 2020:30–44

    work page 2020

  16. [16]

    Practical Cloud Workloads for Serverless FaaS

    Kim J, Lee K. Practical Cloud Workloads for Serverless FaaS. In: ACM. 2019:477–477

    work page 2019

  17. [17]

    Function Compute - Fully Managed Serverless Compute Service

    Alibaba Cloud . Function Compute - Fully Managed Serverless Compute Service. https://www.aliyun.com/product/fc; 2026. [Online; Accessed: 2026-03-24]

    work page 2026

  18. [18]

    INFless: a native serverless system for low-latency, high-throughput inference

    Yang Y , Zhao L, Li Y , et al. INFless: a native serverless system for low-latency, high-throughput inference. In: ACM. 2022:768–781

    work page 2022

  19. [19]

    Serverless computing on heterogeneous computers

    Du D, Liu Q, Jiang X, Xia Y , Zang B, Chen H. Serverless computing on heterogeneous computers. In: ACM. 2022:797–813

    work page 2022

  20. [20]

    {SkyPilot}: An intercloud broker for sky computing

    Yang Z, Wu Z, Luo M, et al. {SkyPilot}: An intercloud broker for sky computing. In: USENIX Association. 2023:437–455

    work page 2023

  21. [21]

    Google Cloud accidentally deletes UniSuper’s online account due to ’unprecedented misconfiguration’.The Guardian; 2024

    Taylor J. Google Cloud accidentally deletes UniSuper’s online account due to ’unprecedented misconfiguration’.The Guardian; 2024. Available at: https://www.theguardian.com/australia-news/article/2024/may/09/unisuper-google-cloud-issue-account-access (Accessed: March 27, 2026)

    work page 2024

  22. [22]

    The inherent mechanism and a case study of the constructional evolution of the jointcloud ecosystem

    Shi P, Liu H, Yang S, Zhang Y , Zhong Y . The inherent mechanism and a case study of the constructional evolution of the jointcloud ecosystem. IEEE Internet of Things Journal.2019;7(3):1561–1571

    work page 2019

  23. [23]

    Sky computing.IEEE Internet Computing.2009;13(5):43–51

    Keahey K, Tsugawa M, Matsunaga A, Fortes J. Sky computing.IEEE Internet Computing.2009;13(5):43–51

    work page 2009

  24. [24]

    Serverless Computing - AWS Lambda - Amazon Web Services

    Amazon Web Services, Inc. . Serverless Computing - AWS Lambda - Amazon Web Services. https://aws.amazon.com/cn/lambda/; 2025. Accessed March 24, 2026

    work page 2025

  25. [25]

    Function Compute FC - Fully Managed Serverless Computing Service - Alibaba Cloud

    Alibaba Cloud . Function Compute FC - Fully Managed Serverless Computing Service - Alibaba Cloud. https://www.aliyun.com/product/fc; 2026. Accessed March 24, 2026

    work page 2026

  26. [26]

    The docker containerization platform,

    Docker . The docker containerization platform,. https://www.docker.com; 2026. Accessed: 2026-03-27

    work page 2026

  27. [27]

    Firecracker: Lightweight virtualization for serverless applications

    Agache A, Brooker M, Iordache A, et al. Firecracker: Lightweight virtualization for serverless applications. In: USENIX Association. 2020:419– 434

    work page 2020

  28. [28]

    Google gvisor: Container runtime sandbox,

    gVisor G. Google gvisor: Container runtime sandbox,. https://github.com/google/gvisor; 2026. Accessed: 2026-03-27

    work page 2026

  29. [29]

    A Measurement Study on Serverless Workflow Services

    Wen J, Liu Y . A Measurement Study on Serverless Workflow Services. In: IEEE. 2021:741–750

    work page 2021

  30. [30]

    Bert: Pre-training of deep bidirectional transformers for language understanding

    Devlin J, Chang MW, Lee K, Toutanova K. Bert: Pre-training of deep bidirectional transformers for language understanding. In: Association for Computational Linguistics. 2019:4171–4186

    work page 2019

  31. [31]

    Deep Residual Learning for Image Recognition

    He K, Zhang X, Ren S, Sun J. Deep Residual Learning for Image Recognition. In: IEEE. 2016:770–778

    work page 2016

  32. [32]

    Shuffling, Fast and Slow: Scalable Analytics on Serverless Infrastructure.Networked Systems Design and Implementation.2019:193–206

    Pu Q, Venkataraman S, Stoica I. Shuffling, Fast and Slow: Scalable Analytics on Serverless Infrastructure.Networked Systems Design and Implementation.2019:193–206

    work page 2019

  33. [33]

    <i>Astrea:</i>Auto-Serverless Analytics Towards Cost-Efficiency and QoS-Awareness.IEEE Transactions on Parallel and Distributed Systems.2022;33:3833–3849

    Jarachanthan J, Chen L, Xu F, Li B. <i>Astrea:</i>Auto-Serverless Analytics Towards Cost-Efficiency and QoS-Awareness.IEEE Transactions on Parallel and Distributed Systems.2022;33:3833–3849. doi: 10.1109/tpds.2022.3172069

    work page doi:10.1109/tpds.2022.3172069 2022

  34. [34]

    Edge-assisted adaptive configuration for serverless-based video analytics.IEEE Transactions on Networking.2025

    Wang Z, Zhang R, Zhang S, Cheng W, Wang W, Cui Y . Edge-assisted adaptive configuration for serverless-based video analytics.IEEE Transactions on Networking.2025

    work page 2025

  35. [35]

    Boki: Stateful serverless computing with shared logs

    Jia Z, Witchel E. Boki: Stateful serverless computing with shared logs. In: ACM. 2021:691–707

    work page 2021

  36. [36]

    FaaSt: Optimize makespan of serverless workflows in federated commercial FaaS

    Ristov S, Gritsch P. FaaSt: Optimize makespan of serverless workflows in federated commercial FaaS. In: IEEE. 2022:183–194

    work page 2022

  37. [37]

    xAFCL: Run Scalable Function Choreographies Across Multiple FaaS Systems.IEEE Transactions on Services Computing.2021:1–1

    Ristov S, Pedratscher S, Fahringer T. xAFCL: Run Scalable Function Choreographies Across Multiple FaaS Systems.IEEE Transactions on Services Computing.2021:1–1. doi: 10.1109/tsc.2021.3128137

    work page doi:10.1109/tsc.2021.3128137 2021

  38. [38]

    XFaaS: Cross-platform Orchestration of FaaS Workflows on Hybrid Clouds

    Khochare A, Khare T, Kulkarni V , Simmhan Y . XFaaS: Cross-platform Orchestration of FaaS Workflows on Hybrid Clouds. In: ACM. 2023:498–512

    work page 2023

  39. [39]

    AWS Step Functions - Visual Workflow Service

    Amazon Web Services . AWS Step Functions - Visual Workflow Service. https://aws.amazon.com/cn/step-functions; 2026. [Online; Accessed: 2026-03-24]

    work page 2026

  40. [40]

    Google cloud composer,

    Google . Google cloud composer,. https://cloud.google.com/composer; 2026. Accessed: 2026-03-27

    work page 2026

  41. [41]

    Durable Functions overview

    Azure . Durable Functions overview. Microsoft Learn; 2026. Available at: https://learn.microsoft.com/en-us/azure/azure-functions/durable/ durable-functions-overview (Accessed: March 27, 2026)

    work page 2026

  42. [42]

    CloudFlow: Serverless Application Orchestration for DevOps

    Alibaba Cloud . CloudFlow: Serverless Application Orchestration for DevOps. https://www.alibabacloud.com/en/product/serverless-workflow

    work page

  43. [43]

    Accessed: 2026-03-24

    work page 2026

  44. [44]

    Cloudburst

    Sreekanti V , Wu C, Lin XC, et al. Cloudburst. In: . 13. VLDB Endowment. 2020:2438–2452

    work page 2020

  45. [45]

    Wukong: A scalable and locality-enhanced framework for serverless parallel computing

    Carver B, Zhang J, Wang A, Anwar A, Wu P, Cheng Y . Wukong: A scalable and locality-enhanced framework for serverless parallel computing. In: ACM. 2020:1–15

    work page 2020

  46. [46]

    Netherite: Efficient execution of serverless workflows.Proceedings of the VLDB Endowment

    Burckhardt S, Chandramouli B, Gillum C, et al. Netherite: Efficient execution of serverless workflows.Proceedings of the VLDB Endowment. 2022;15(8):1591–1604

    work page 2022

  47. [47]

    Occupy the Cloud: Distributed Computing for the 99%

    Jonas E, Pu Q, Venkataraman S, Stoica I, Recht B. Occupy the Cloud: Distributed Computing for the 99doi: 10.48550/arxiv.1702.04024

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.1702.04024

  48. [48]

    Serverless Data Analytics in the IBM Cloud

    Sampé J, Vernik G, Sánchez-Artigas M, García-López P. Serverless Data Analytics in the IBM Cloud. In: N. A. 2018:1–8

    work page 2018

  49. [49]

    Amazon Web Services (AWS)

    Amazon Web Services . Amazon Web Services (AWS). https://aws.amazon.com/; 2026. Accessed: 2026-03-27

    work page 2026

  50. [50]

    Google Cloud Platform (GCP)

    Google Cloud . Google Cloud Platform (GCP). https://cloud.google.com/; 2026. Accessed: 2026-03-27

    work page 2026

  51. [51]

    Realizing the {Fault-Tolerance} promise of cloud storage using locks with intent

    Setty S, Su C, Lorch JR, et al. Realizing the {Fault-Tolerance} promise of cloud storage using locks with intent. In: USENIX Association. 2016:501–516

    work page 2016

  52. [52]

    HyperFlow: A model of computation, programming approach and enactment engine for complex distributed workflows.Future Generation Computer Systems.2015;55:147–162

    Bali´s B. HyperFlow: A model of computation, programming approach and enactment engine for complex distributed workflows.Future Generation Computer Systems.2015;55:147–162. doi: 10.1016/j.future.2015.08.015

    work page doi:10.1016/j.future.2015.08.015 2015

  53. [53]

    Lithops,

    cloud l. Lithops,. https://github.com/lithops-cloud/lithops; 2026. Accessed: 2026-03-27

    work page 2026

  54. [54]

    GlobalFlow: A Cross-Region Orchestration Service for Serverless Computing Services

    Ge Z, Peng Y . GlobalFlow: A Cross-Region Orchestration Service for Serverless Computing Services. In: IEEE. 2019 Jointλ 21

    work page 2019

  55. [55]

    faas-sim: A trace-driven simulation framework for serverless edge computing platforms.Software: Practice and experience.2023;53(12):2327–2361

    Raith P, Rausch T, Furutanpey A, Dustdar S. faas-sim: A trace-driven simulation framework for serverless edge computing platforms.Software: Practice and experience.2023;53(12):2327–2361

    work page 2023

  56. [56]

    RADF: Architecture decomposition for function as a service.Software: Practice and Experience.2024;54(4):566– 594

    Zhu L, Tamburri DA, Casale G. RADF: Architecture decomposition for function as a service.Software: Practice and Experience.2024;54(4):566– 594

    work page 2024

  57. [57]

    Functions as a service for distributed deep neural network inference over the cloud-to-things continuum

    Bueno A, Rubio B, Martín C, Díaz M. Functions as a service for distributed deep neural network inference over the cloud-to-things continuum. Software: Practice and Experience.2024;54(8):1297–1311

    work page 2024

  58. [58]

    Amazon S3: Cloud Object Storage

    Amazon Web Services . Amazon S3: Cloud Object Storage. https://aws.amazon.com/s3/; 2026. Accessed: 2026-03-24

    work page 2026

  59. [59]

    Object Storage Service (OSS): Secure, Reliable, and Low-cost Cloud Storage

    Alibaba Cloud . Object Storage Service (OSS): Secure, Reliable, and Low-cost Cloud Storage. https://www.alibabacloud.com/en/product/ object-storage-service; 2026. Accessed: 2026-03-24

    work page 2026

  60. [60]

    Amazon DynamoDB Serverless Database - AWS Cloud Services

    Amazon Web Services, Inc. . Amazon DynamoDB Serverless Database - AWS Cloud Services. https://aws.amazon.com/cn/dynamodb/; 2025. Accessed March 24, 2026

    work page 2025

  61. [61]

    Tablestore - Massive Structured Data Storage, Retrieval, and Analysis - Alibaba Cloud

    Alibaba Cloud . Tablestore - Massive Structured Data Storage, Retrieval, and Analysis - Alibaba Cloud. https://www.aliyun.com/product/ots; 2026. Accessed March 24, 2026

    work page 2026

  62. [62]

    Primula: A practical shuffle/sort operator for serverless computing

    Sánchez-Artigas M, Eizaguirre GT, Vernik G, Stuart L, García-López P. Primula: A practical shuffle/sort operator for serverless computing. In: ACM. 2020:31–37

    work page 2020

  63. [63]

    {ORION} and the three rights: Sizing, bundling, and prewarming for serverless {DAGs}

    Mahgoub A, Yi EB, Shankar K, Elnikety S, Chaterji S, Bagchi S. {ORION} and the three rights: Sizing, bundling, and prewarming for serverless {DAGs}. In: USENIX Association. 2022:303–320

    work page 2022

  64. [64]

    serverless-bert-huggingface-aws-lambda-docker,

    Schmid P. serverless-bert-huggingface-aws-lambda-docker,. https://github.com/philschmid/serverless-bert-huggingface-aws-lambda-docker; 2026. Accessed: 2026-03-27

    work page 2026

  65. [65]

    Google . Gemini. https://gemini.google.com/app; 2026. Accessed: 2026-03-28

    work page 2026