Leveraging Network Methods for Hub-like Microservice Detection

Alexander Bakhtin; Davide Taibi; Matteo Esposito; Valentina Lenarduzzi

REVIEW 3 major objections 5 minor 56 references

Leveraging Network Methods for Hub-like Microservice Detection

T0 review · 3 major / 5 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read The non-parametric Erdős–Rényi encoding method by Kirkley, applied to incoming connections, detects Hub-like microservices with the highest precision among the considered methods on 25 service dependency graphs.

desk verdict A useful first comparison of MDL-based hub detection for microservices, with a plausible ER-in recommendation that is only as solid as the authors' own labels. read the letter →

arxiv 2506.07683 v2 pith:43QFX3NP submitted 2025-06-09 cs.SE cs.DM

Alexander Bakhtin , Matteo Esposito , Valentina Lenarduzzi , Davide Taibi This is my paper

classification cs.SEcs.DM

keywords microservicearchitecturehub-likeanti-patternservicedependencygraphnetworkhubdetectionminimumdescriptionlengthErdős–Rényiencodingdegreecentralityscale-freenetworks

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

This paper asks which network-science method can reliably detect the Hub-like microservice anti-pattern—a service with unusually many dependencies—and evaluates several families of detectors on 25 service dependency graphs with 251 services. It finds that the studied microservice networks are not scale-free, that the various detectors disagree sharply about which services are hubs, and that the most accurate detector is Kirkley's minimum-description-length method using the Erdős–Rényi encoding on incoming connections (ER-in). With manual validation as ground truth, ER-in flags 33 hubs at precision 0.848 when infrastructural hubs count as true positives, and still 0.783 when they are ignored. The paper concludes that practitioners should prefer degree centrality over raw degree and should consider adopting ER-in in tools like Arcan.

What carries the argument

The method carrying the argument is Kirkley's minimum-description-length hub detection under the Erdős–Rényi (ER) encoding. Given a directed network, the algorithm seeks the set of hub nodes whose removal and separate description minimizes the total number of bits needed to transmit the network, relying on the fact that compressing a network is cheaper when a few high-degree nodes are described first; it is fully non-parametric, requiring no threshold. The paper applies it to incoming and outgoing connections separately (ER-in and ER-out), and compares it against centrality metrics combined with the clustering coefficient, scale-free power-law fitting, and the quartile-based thresholding of the Arcan tool.

What would settle it

Have a panel of architects, blind to method outputs, independently label the same 251 services in the 25 SDGs; if their Hub-like labels disagree substantially with the authors' 44 hubs, the precision values that rank ER-in first would not be reproducible. Alternatively, run ER-in on a large industrial microservice system with known architectural reviews and check whether its detections match services an architect would actually inspect.

Watch

Extended reading notes

Core claim

The central claim is that, among the hub detection approaches considered, the non-parametric ER-in method—Alec Kirkley's algorithm that labels the set of hubs minimizing the bits needed to describe the network under an Erdős–Rényi model—is the most accurate and practical detector of Hub-like microservices. On the authors' manually labeled ground truth, ER-in detected 33 hubs with precision 0.848 (infrastructural hubs counted as true positives), 0.783 (infrastructural hubs ignored), and 0.545 (infrastructural hubs counted as false positives), outperforming all other methods while returning a manageable number of candidates. The paper also establishes that the 25 networks are not scale-free at the 0.01 significance level, that most methods do not agree with one another (Fleiss' kappa 0.0323 across all detectors), and that the Arcan tool's absolute-degree threshold detects hubs in only one system while its normalized degree-centrality variant detects hubs in 22 systems, mostly infrastructural ones.

Load-bearing premise

The precision ranking stands on the authors' manual labels of which of the 251 services are Hub-like, so if other architects would label the same services differently, the ranking could change.

Editorial extensions

If this is right

Tools such as Arcan could be updated to use normalized degree centrality or the ER-in algorithm instead of absolute degree thresholds, yielding detections across many systems rather than only the largest one.
Practitioners should treat detected hubs from most centrality-based methods cautiously, since these methods mostly returned infrastructural hubs in this dataset.
The absence of scale-free structure in the studied SDGs rules out explanations of hubs as emergent properties of preferential attachment in these systems.
Because ER-in considers incoming connections, the results suggest that in these systems the hubs are services that call many others rather than services called by many.
A usable benchmark for Hub-like detection should separate infrastructural hubs from business-logic hubs, since no method naturally filters them out.

Reading between the lines

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

Not stated in the paper: if ER-in's edge stems from how closely SDGs resemble random graphs, the method could transfer to other dependency networks such as package imports or data-flow graphs.
Not stated in the paper: an architect-blind re-labeling of the same 25 SDGs would test whether the 44-hub ground truth is stable enough to support the precision ranking.
Not stated in the paper: whitelisting infrastructure services before detection would likely raise ER-in's precision further and make the method more directly usable in practice.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Desk Editor's Note

read the letter

Colleague, this paper is the first to try Kirkley's minimum-description-length hub detection on microservice dependency graphs, and it does a creditably systematic job of comparing it against centrality, clustering, and the Arcan threshold approach. The replication package is real, and they are refreshingly open about the trickiest issue: infrastructural hubs like eureka and consul inflate detection counts and dominate many methods. The recommendation to prefer normalized degree centrality over absolute degree is well supported by the data — Arcan abs detects almost nothing outside train-ticket, while Arcan norm finds candidates across most systems.

The soft spots are real, but mostly the ones the authors themselves flag. The ground truth is the authors' own manual labeling of 44 hubs and 51 infrastructural hubs, and they do not report that the labeling was blind to which method detected each service. That is a genuine threat: if the raters expected certain services to be hubs, the precision ranking could shift. The paper reports three precision variants (IH as TP, ignored, FP), but the winner ER-in still looks good under the first two; under the most hostile variant, its 0.545 is only slightly behind Arcan abs's 0.6, which detected only ten services in two systems, so I would not call the central claim fragile on that score. Still, there are no confidence intervals on any precision estimate, and with 33 detected hubs, sampling noise is real.

The scale-free result is weaker than the abstract suggests. The p-values for degree, in-degree, and out-degree are 0.035, 0.053, and 0.055. Using alpha=0.01 makes them conclude 'not scale-free,' but that is a conservative threshold and they don't report the power-law package's comparisons against alternative distributions, so I would read the scale-free claim as 'not detectable in these small SDGs' rather than a strong negative.

One more note: the centrality-family hub lists depend on the two authors manually picking clusters in scatter plots. That is essentially a manual detector, so its agreement with other methods is not a fair benchmark.

Overall: a useful, honest, within-subfield contribution. The main value is the comparison and the practical argument for normalization, and the ER-in finding is plausible but needs replication on larger, industrial systems or at least an independent rater study. I would send it to peer review with a request to add confidence intervals and to describe the labeling procedure in enough detail to judge potential bias, but the work deserves referee time.

Referee Report

3 major / 5 minor

Summary. The paper compares network-based hub detection techniques for identifying the Hub-like microservice anti-pattern in 25 service dependency graphs (SDGs) reconstructed from open-source microservice systems. The evaluated techniques include scale-free property testing, centrality/clustering combinations, Kirkley's minimum-description-length methods (ER and CM encodings), and a replication of the Arcan tool's threshold-based approach. The authors report that the studied networks are not scale-free, that most methods disagree on detected hubs, and that the ER-in method achieves the best trade-off between number of detections and precision (33 hubs, precision 0.848 when infrastructural hubs are counted as true positives). The paper includes a replication package and explicitly discusses threats to validity.

Significance. If the results hold, the paper provides one of the first systematic comparisons of network-science methods for detecting Hub-like microservices, a notoriously subjectively defined anti-pattern. The study is transparent about its limitations, ships a replication package, and leverages existing implementations (e.g., Kirkley's code) rather than introducing ad-hoc methods. The recommendation to prefer normalized degree centrality over absolute degree, and to consider the ER-in method, is potentially useful for practitioners. However, the empirical evidence for the central claim rests on a subjective, author-created ground truth and on a particular treatment of infrastructural hubs, so the comparative ranking should be regarded as preliminary until independently validated.

major comments (3)

[Section V-F, Table IV] The central claim that ER-in is the most accurate hub detection method is sensitive to the chosen treatment of infrastructural hubs. In Table IV, when infrastructural hubs are counted as false positives, Arcan abs achieves precision 0.6 against ER-in's 0.545, and when they are ignored, Arcan abs shows precision 1.0 with 10 detections. The paper asserts that 'the best method according to precision and number of detections is ER in' without a statistical test or a principled justification for choosing the 'IH as TP' or 'IH ignored' variants over the 'IH as FP' variant. Because this ranking supports the main recommendation in Sections VI and VIII, the authors should provide confidence intervals, a sensitivity analysis over the IH treatment, and a clearer rationale tied to the practitioner scenario they target.
[Section IV-D-6 and Section V-F] The precision ground truth consists of labels assigned by the two authors, who inspected the services detected by each method. The paper does not state that the labeling was performed blind to the detecting method, and the same authors manually selected candidates for the centrality/clustering family (Section IV-D-3). This creates a risk that the labels are systematically influenced by expectations about which methods should perform well, and it means the precision ranking is not independent of the authors' prior beliefs. The reported Cohen's Kappa demonstrates inter-rater agreement between the two authors but not independence from the authors' perspective. The authors should either recruit independent raters (e.g., practitioners or other researchers) or clearly scope the conclusions as an author-consensus evaluation and temper the recommendations accordingly.
[Section V-F and Section VI] The three-way treatment of infrastructural hubs (TP, FP, or ignored) changes the relative ordering of methods, and the paper does not assess the impact of whitelisting or removing infrastructural hubs from the networks before applying the methods. Since the authors themselves suggest that infrastructural hubs 'should either be removed from the network before analysis or whitelisted' (Section VI), the precision comparison should include an analysis after such pre-processing for all methods. Without this, the practical recommendation to practitioners remains ambiguous, as the reported precision numbers are conditional on an arbitrary decision about how to treat IH.

minor comments (5)

[Section IV-D-1] The powerlaw package's p-value thresholds are applied separately to degree, in-degree, and out-degree without correcting for multiple testing; this is a minor statistical caveat that could be noted.
[Figure 5] The label 'EM_in' appears to be a typo for 'ER_in'; please ensure consistent notation throughout the tables and figures.
[Section III] The text refers to 'Designate' but the tool is 'Designite'; please fix this typo.
[Table IV] The row for CM out shows dashes for all precision columns; consider adding a sentence explaining that this method detected no hubs.
[Section VI] The phrase 'almost scale-free' is a strong interpretation of p-values 0.035, 0.053, and 0.055; consider rewording to avoid overstatement given the small dataset.

Circularity Check

0 steps flagged · score 0.0 of 10

No derivation-level circularity: the hub-detection methods are external and no fitted parameter is renamed as a prediction.

full rationale

The paper's central claim is an empirical benchmark: several external hub-detection algorithms (Kirkley's ER/CM MDL encodings and baselines, centrality+clustering combinations, Arcan replications) are applied to 25 SDGs and their outputs are compared to a manually labeled set of 44 hubs and 51 infrastructural hubs (Section IV-D-6, Section V-F). No method is fitted to those labels: Kirkley's code is parameter-free and published externally, and the Arcan replications use the original quartile-based threshold procedure rather than a threshold tuned to the validation labels. The precision values in Table IV are by definition the overlap of each method's output with the authors' labels, so the ranking is a measurement, not a derivation that reduces to its inputs. The self-referential element that exists is evaluative: the same two authors who selected centrality/clustering candidates (Section IV-D-3) also assigned the Hub-like labels, and the paper itself flags this in Section VII ('the results of the validation might be biased towards the researchers' point of view'). This is a construct-validity threat about label reliability, not a circularity in the derivation chain; no equation, definition, or self-citation forces the ER-in result. The dataset is taken from the authors' prior work [7], but a dataset is an external artifact and its use is not a load-bearing self-citation. No circular step meeting the quote-and-reduction threshold was found.

Assumptions & free parameters 0 free parameters · 4 assumptions · 0 invented entities

The paper introduces no new theoretical entities; it imports existing network methods and applies them to service dependency graphs. The only parameters are those inherent to the methods being evaluated (e.g., Arcan's quartile threshold), not new fitted constants.

assumptions (4)

domain assumption The Hub-like anti-pattern definition of Cerny et al. (2023) is the operational definition.
The entire study operationalizes the anti-pattern as a service with many outgoing and ingoing dependencies; all manual labels follow this definition.
domain assumption Code2DFD static reconstruction yields accurate service dependency graphs.
The 24 SDGs from Bakhtin et al. [7] are assumed to faithfully represent inter-service calls.
domain assumption Manual validation by two authors provides a reliable ground truth for Hub-likeness.
Precision is computed against author-generated labels; the authors acknowledge this is subjective and potentially biased in Section VII.
domain assumption Power-law significance testing with the powerlaw package is appropriate for these network sizes.
The conclusion that networks are not scale-free rests on goodness-of-fit p-values, which are sensitive to small samples; the authors note this in Section VI.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Leveraging Network Methods for Hub-like Microservice Detection." pith.science (2026). https://pith.science/paper/43QFX3NP

@misc{pith2026250607683,
  author       = {Pith},
  title        = {Pith review of: Leveraging Network Methods for Hub-like Microservice Detection},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/43QFX3NP}},
  note         = {Machine review of arXiv:2506.07683}
}

read the original abstract

Context: Microservice Architecture is a popular architectural paradigm that facilitates flexibility by decomposing applications into small, independently deployable services. Catalogs of architectural anti-patterns have been proposed to highlight the negative aspects of flawed microservice design. In particular, the Hub-like anti-pattern lacks an unambiguous definition and detection method. Aim: In this work, we aim to find a robust detection approach for the Hub-like microservice anti-pattern that outputs a reasonable number of Hub-like candidates with high precision. Method: We leveraged a dataset of 25 microservice networks and several network hub detection techniques to identify the Hub-like anti-pattern, namely scale-free property, centrality metrics and clustering coefficient, minimum description length principle, and the approach behind the Arcan tool. Results and Conclusion: Our findings revealed that the studied architectural networks are not scale-free, that most considered hub detection approaches do not agree on the detected hubs, and that the method by Kirkley leveraging the Erdos-Renyi encoding is the most accurate one in terms of the number of detected hubs and the detection precision. Investigating further the applicability of these methods to detecting Hub-like components in microservice-based and other systems opens up new research directions. Moreover, our results provide an evaluation of the approach utilized by the widely used Arcan tool and highlight the potential to update the tool to use the normalized degree centrality of a component in the network, or for the approach based on ER encoding to be adopted instead.

Figures

Figures reproduced from arXiv: 2506.07683 by the authors.

**Figure 2.** Data collection and analysis process use the Designate method, since it uses an absolute threshold of the Fan-in and Fan-out of a class which cannot be migrated to the inter-service level, while the ARCADE method is very similar to the considered Loubar method, while suffering from the concerns mentioned in Section III. RQ2 What is the agreement of detected Hub-like microservices among the scale-free, network-based… view at source ↗

**Figure 3.** Power law fits of the distributions of node degrees to determine the scale-free property [PITH_FULL_IMAGE:figures/full_fig_p008_3.png] view at source ↗

Figures from the paper (2 more)

**Figure 6.** Figure 6: Jaccard index matrix for methods considering outgoing connections [PITH_FULL_IMAGE:figures/full_fig_p008_6.png]

**Figure 7.** Figure 7: Jaccard index matrix for methods considering both the incoming and [PITH_FULL_IMAGE:figures/full_fig_p008_7.png]

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

56 extracted references · 55 canonical work pages

[1]

Microservices: a definition of this new architectural term,

J. Lewis and M. Fowler, “Microservices: a definition of this new architectural term,” 2014. [Online]. Available: https://martinfowler.com/ articles/microservices.html

work page 2014
[2]

Survey on tools and techniques detecting microservice api patterns,

A. Bakhtin, A. Al Maruf, T. Cerny, and D. Taibi, “Survey on tools and techniques detecting microservice api patterns,” in 2022 IEEE International Conference on Services Computing (SCC) . IEEE, 2022, pp. 31–38

work page 2022
[3]

Catalog and detection techniques of microservice anti-patterns and bad smells: A tertiary study,

T. Cerny, A. S. Abdelfattah, A. Al Maruf, A. Janes, and D. Taibi, “Catalog and detection techniques of microservice anti-patterns and bad smells: A tertiary study,” Journal of Systems and Software , vol. 206, p. 111829, 2023

work page 2023
[4]

Towards microservice smells detection,

I. Pigazzini, F. A. Fontana, V . Lenarduzzi, and D. Taibi, “Towards microservice smells detection,” in Proceedings of the 3rd International Conference on Technical Debt , 2020, pp. 92–97

work page 2020
[5]

Using microservice telemetry data for system dynamic analysis,

A. Al Maruf, A. Bakhtin, T. Cerny, and D. Taibi, “Using microservice telemetry data for system dynamic analysis,” in 2022 IEEE International Conference on Service-Oriented System Engineering (SOSE) . IEEE, 2022, pp. 29–38

work page 2022
[6]

Microservice api pattern detection : Using business pro- cesses and call graphs,

A. Bakhtin, “Microservice api pattern detection : Using business pro- cesses and call graphs,” Master’s thesis, Tampere University, 2022

work page 2022
[7]

Network centrality as a new perspective on microservice architecture,

A. Bakhtin, M. Esposito, V . Lenarduzzi, and D. Taibi, “Network centrality as a new perspective on microservice architecture,” in 2025 IEEE 22nd International Conference on Software Architecture (ICSA) , 2025, pp. 72–83

work page 2025
[8]

Identifying hubs in directed networks,

A. Kirkley, “Identifying hubs in directed networks,” Phys. Rev. E , vol. 109, p. 034310, Mar 2024

work page 2024

Show all 56 references

[9]

Architectural smells detected by tools: a catalogue proposal,

U. Azadi, F. A. Fontana, and D. Taibi, “Architectural smells detected by tools: a catalogue proposal,” in 2019 IEEE/ACM International Conference on Technical Debt (TechDebt) . IEEE, 2019, pp. 88–97

2019
[10]

Scale-free networks,

A.-L. Barab ´asi and E. Bonabeau, “Scale-free networks,” Scientific amer- ican, vol. 288, no. 5, pp. 50–9, 2003

2003
[11]

Scale-free networks: a decade and beyond,

A.-L. Barab ´asi, “Scale-free networks: a decade and beyond,” science, vol. 325, no. 5939, pp. 412–413, 2009

2009
[12]

Automatic met- ric thresholds derivation for code smell detection,

F. A. Fontana, V . Ferme, M. Zanoni, and A. Yamashita, “Automatic met- ric thresholds derivation for code smell detection,” in 2015 IEEE/ACM 6th International Workshop on Emerging Trends in Software Metrics . IEEE, 2015, pp. 44–53

2015
[13]

Arcan: A tool for architectural smells detection,

F. A. Fontana, I. Pigazzini, R. Roveda, D. Tamburri, M. Zanoni, and E. Di Nitto, “Arcan: A tool for architectural smells detection,” in 2017 IEEE International Conference on Software Architecture Workshops (ICSAW). IEEE, 2017, pp. 282–285

2017
[14]

A study on the taxonomy of service antipat- terns,

F. Palma and N. Mohay, “A study on the taxonomy of service antipat- terns,” in 2015 IEEE 2nd International Workshop on Patterns Promotion and Anti-patterns Prevention (PPAP) . IEEE, 2015, pp. 5–8

2015
[15]

Automated code-smell detection in microservices through static analysis: A case study,

A. Walker, D. Das, and T. Cerny, “Automated code-smell detection in microservices through static analysis: A case study,” Applied Sciences, vol. 10, no. 21, p. 7800, 2020

2020
[16]

Automated microservice code-smell detection,

——, “Automated microservice code-smell detection,” in Information Science and Applications: Proceedings of ICISA 2020 . Springer, 2021, pp. 211–221

2020
[17]

On microservice analysis and architecture evolution: A systematic mapping study,

V . Bushong, A. S. Abdelfattah, A. A. Maruf, D. Das, A. Lehman, E. Jaroszewski, M. Coffey, T. Cerny, K. Frajtak, P. Tisnovskyet al., “On microservice analysis and architecture evolution: A systematic mapping study,” Applied Sciences, vol. 11, no. 17, p. 7856, 2021

2021
[18]

Technical debt resulting from architectural degradation and code smells: a systematic mapping study,

D. Das, A. A. Maruf, R. Islam, N. Lambaria, S. Kim, A. S. Abdelfattah, T. Cerny, K. Frajtak, M. Bures, and P. Tisnovsky, “Technical debt resulting from architectural degradation and code smells: a systematic mapping study,” ACM SIGAPP Applied Computing Review , vol. 21, no. 4,...

2022
[19]

Microvision: Static analysis-based approach to visualizing microser- vices in augmented reality,

T. Cerny, A. S. Abdelfattah, V . Bushong, A. Al Maruf, and D. Taibi, “Microvision: Static analysis-based approach to visualizing microser- vices in augmented reality,” in 2022 IEEE International Conference on Service-Oriented System Engineering (SOSE) . IEEE, 2022, pp. 49–58

2022
[20]

Detecting microservice anti-patterns using interactive service call graphs: Effort assessment,

A. Huizinga, G. Parker, A. S. Abdelfattah, X. Li, T. Cerny, and D. Taibi, “Detecting microservice anti-patterns using interactive service call graphs: Effort assessment,” in Southwest Data Science Conference . Springer, 2023, pp. 212–227

2023
[21]

Visualizing microservice architecture in the dynamic perspective: A systematic mapping study,

M. E. Gortney, P. E. Harris, T. Cerny, A. Al Maruf, M. Bures, D. Taibi, and P. Tisnovsky, “Visualizing microservice architecture in the dynamic perspective: A systematic mapping study,” IEEE Access , vol. 10, pp. 119 999–120 012, 2022

2022
[22]

From static code analysis to visual models of microservice architecture,

T. Cerny, A. S. Abdelfattah, J. Yero, and D. Taibi, “From static code analysis to visual models of microservice architecture,” Cluster Computing, vol. 27, no. 4, pp. 4145–4170, 2024

2024
[23]

Designite: A software design quality assessment tool,

T. Sharma, P. Mishra, and R. Tiwari, “Designite: A software design quality assessment tool,” in Proceedings of the 1st international work- shop on bringing architectural design thinking into developers’ daily activities, 2016, pp. 1–4

2016
[24]

Relating architectural decay and sustainability of software systems,

D. M. Le, C. Carrillo, R. Capilla, and N. Medvidovic, “Relating architectural decay and sustainability of software systems,” in 2016 13th Working IEEE/IFIP Conference on Software Architecture (WICSA). IEEE, 2016, pp. 178–181

2016
[25]

Detecting architecture debt in micro-service open-source projects,

R. Capilla, F. A. Fontana, T. Mikkonen, P. Bacchiega, and V . Salamanca, “Detecting architecture debt in micro-service open-source projects,” in 2023 49th Euromicro Conference on Software Engineering and Advanced Applications (SEAA) , 2023, pp. 394–401

2023
[26]

Refactoring of a microservices project driven by architectural smell detection,

P. Bacchiega, D. Rusconi, P. Mereghetti, and F. A. Fontana, “Refactoring of a microservices project driven by architectural smell detection,” in 2024 IEEE 21st International Conference on Software Architecture Companion (ICSA-C). IEEE, 2024, pp. 281–288

2024
[27]

Exploiting dynamic analysis for architectural smell detection: a preliminary study,

I. Pigazzini, D. Di Nucci, F. A. Fontana, and M. Belotti, “Exploiting dynamic analysis for architectural smell detection: a preliminary study,” in 2022 48th Euromicro Conference on Software Engineering and Advanced Applications (SEAA) . IEEE, 2022, pp. 282–289

2022
[28]

Msvision: An extensible tool for detecting patterns and bad smells in heterogeneous microservices systems,

J. F. Lino Daniel, M. Vidu, T. Rosa, A. Goldman, and E. Guerra, “Msvision: An extensible tool for detecting patterns and bad smells in heterogeneous microservices systems,” Available at SSRN 4892088 , 2024

2024
[29]

A catalog of micro frontends anti- patterns,

N. Silva, E. Rodrigues, and T. Conte, “A catalog of micro frontends anti- patterns,” 2024. [Online]. Available: https://arxiv.org/abs/2411.19472

2024 arXiv
[30]

Using Graph Theory and Network Science to Explore Your Microservices Architecture,

N. Watt, “Using Graph Theory and Network Science to Explore Your Microservices Architecture,” in Proceedings of GOTO Berlin 2019 , 2019

2019
[31]

Evaluation of microservice architectures: A metric and tool-based approach,

T. Engel, M. Langermeier, B. Bauer, and A. Hofmann, “Evaluation of microservice architectures: A metric and tool-based approach,” in Information Systems in the Big Data Era: CAiSE Forum 2018, Tallinn, Estonia, June 11-15, 2018, Proceedings 30. Springer, 2018, pp. 74–89

2018
[32]

Service dependency graph analysis in microservice architecture,

E. Gaidels and M. Kirikova, “Service dependency graph analysis in microservice architecture,” in Perspectives in Business Informatics Research: 19th International Conference on Business Informatics Re- search, BIR 2020, Vienna, Austria, September 21–23, 2020, Proceedings

2020
[33]

Springer International Publishing, 2020, pp. 128–139

2020
[34]

Industry 4.0 and Microservices with the power of Graphs,

H. Farsi, “Industry 4.0 and Microservices with the power of Graphs,” International Workshop of Services and Industry of the Future (IWSIF 2022), Oct. 2022, poster. [Online]. Available: https: //hal.science/hal-03825330

2022
[35]

Algorithm for detecting antipatterns in mi- croservices projects,

R. Borges and T. Khan, “Algorithm for detecting antipatterns in mi- croservices projects,” CEUR workshop proceedings , 2019

2019
[36]

Product metrics for service-oriented infrastructures,

D. Rud, A. Schmietendorf, and R. R. Dumke, “Product metrics for service-oriented infrastructures,” IWSM/MetriKon, pp. 161–174, 2006

2006
[37]

Diameter of the world-wide web,

R. Albert, H. Jeong, and A.-L. Barab ´asi, “Diameter of the world-wide web,” nature, vol. 401, no. 6749, pp. 130–131, 1999

1999
[38]

Identifying thresholds for software fault- iness via optimistic and pessimistic estimations,

L. Lavazza and S. Morasca, “Identifying thresholds for software fault- iness via optimistic and pessimistic estimations,” in Proceedings of the 10th ACM/IEEE International Symposium on Empirical Software Engineering and Measurement , 2016, pp. 1–10

2016
[39]

Slope-based fault-proneness thresholds for software engineering measures,

S. Morasca and L. Lavazza, “Slope-based fault-proneness thresholds for software engineering measures,” in Proceedings of the 20th international conference on evaluation and assessment in software engineering , 2016, pp. 1–10

2016
[40]

Centrality change proneness: an early indicator of microservice architectural degradation,

A. Bakhtin, M. Esposito, V . Lenarduzzi, and D. Taibi, “Centrality change proneness: an early indicator of microservice architectural degradation,”
[41]

Enhancing the defectiveness prediction of methods and classes via jit,

D. Falessi, S. M. Laureani, J. C ¸ arka, M. Esposito, and D. A. d. Costa, “Enhancing the defectiveness prediction of methods and classes via jit,” Empirical Software Engineering , vol. 28, no. 2, p. 37, 2023

2023
[42]

Can we trust the default vulnerabilities severity?

M. Esposito, S. Moreschini, V . Lenarduzzi, D. H¨astbacka, and D. Falessi, “Can we trust the default vulnerabilities severity?” in 2023 IEEE 23rd International Working Conference on Source Code Analysis and Manipulation (SCAM). IEEE, 2023, pp. 265–270

2023
[43]

Clustering and preferential attachment in growing networks,

M. E. Newman, “Clustering and preferential attachment in growing networks,” Physical review E , vol. 64, no. 2, p. 025102, 2001

2001
[44]

Wohlin, P

C. Wohlin, P. Runeson, M. H ¨ost et al. , Experimentation in Software Engineering. Springer, 2012

2012
[45]

Comparison of static analysis architec- ture recovery tools for microservice applications,

S. Schneider, A. Bakhtin, X. Li, J. Soldani, A. Brogi, T. Cerny, R. Scandariato, and D. Taibi, “Comparison of static analysis architec- ture recovery tools for microservice applications,” Empirical Software Engineering, 2025

2025
[46]

Automatic extraction of security-rich dataflow diagrams for microservice applications written in java,

S. Schneider and R. Scandariato, “Automatic extraction of security-rich dataflow diagrams for microservice applications written in java,” Journal of Systems and Software , vol. 202, p. 111722, 2023

2023
[47]

microsecend: A dataset of security-enriched dataflow diagrams for microservice appli- cations,

S. Schneider, T. ¨Ozen, M. Chen, and R. Scandariato, “microsecend: A dataset of security-enriched dataflow diagrams for microservice appli- cations,” in 2023 IEEE/ACM 20th International Conference on Mining Software Repositories (MSR) . IEEE, 2023, pp. 125–129

2023
[48]

A curated dataset of microservices-based systems,

M. I. Rahman, S. Panichella, and D. Taibi, “A curated dataset of microservices-based systems,” arXiv preprint arXiv:1909.03249 , 2019

1909 arXiv
[49]

A feature dataset of microservices-based systems,

W. Yang, B. Song, Y . Xing, Y . Lyu, H. Cui, Z. Liang, and Z. Tu, “A feature dataset of microservices-based systems,” in International Conference on Service Science . Springer, 2024, pp. 73–87

2024
[50]

A dataset of microservices- based open-source projects,

D. A. d’Aragona, A. Bakhtin, X. Li, R. Su, L. Adams, E. Aponte, F. Boyle, P. Boyle, R. Koerner, J. Leeet al., “A dataset of microservices- based open-source projects,” in Proceedings of the 21st International Conference on Mining Software Repositories , 2024, pp. 504–509

2024
[51]

Maximal planar networks with large clustering coefficient and power-law degree distribution,

T. Zhou, G. Yan, and B.-H. Wang, “Maximal planar networks with large clustering coefficient and power-law degree distribution,” Physical Review E—Statistical, Nonlinear, and Soft Matter Physics, vol. 71, no. 4, p. 046141, 2005

2005
[52]

Replication package for “leveraging network methods for hub-like microservice detection

A. Bakhtin, M. Esposito, V . Lenarduzzi, and D. Taibi, “Replication package for “leveraging network methods for hub-like microservice detection”,” May 2025. [Online]. Available: https://zenodo.org/doi/10. 5281/zenodo.15362705

2025
[53]

Fleiss’ kappa statistic without paradoxes,

R. Falotico and P. Quatto, “Fleiss’ kappa statistic without paradoxes,” Quality & Quantity , vol. 49, pp. 463–470, 2015

2015
[54]

A coefficient of agreement for nominal scales,

J. Cohen, “A coefficient of agreement for nominal scales,” Educational and Psychological Measurement , vol. 20, no. 1, pp. 37–46, 1960, publisher: SAGE Publications

1960
[56]

On using network science in mining developers collaboration in software engineering: A systematic literature review,

M. Abufouda and H. Abukwaik, “On using network science in mining developers collaboration in software engineering: A systematic literature review,” International Journal of Data Mining & Knowledge Manage- ment Process (IJDKP) , vol. 7, no. 5/6, pp. 17–34, November 2017

2017
[2025]

Available: https://arxiv.org/abs/2506.07690

[Online]. Available: https://arxiv.org/abs/2506.07690

arXiv

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.

[1] [1]

Microservices: a definition of this new architectural term,

J. Lewis and M. Fowler, “Microservices: a definition of this new architectural term,” 2014. [Online]. Available: https://martinfowler.com/ articles/microservices.html

work page 2014

[2] [2]

Survey on tools and techniques detecting microservice api patterns,

A. Bakhtin, A. Al Maruf, T. Cerny, and D. Taibi, “Survey on tools and techniques detecting microservice api patterns,” in 2022 IEEE International Conference on Services Computing (SCC) . IEEE, 2022, pp. 31–38

work page 2022

[3] [3]

Catalog and detection techniques of microservice anti-patterns and bad smells: A tertiary study,

T. Cerny, A. S. Abdelfattah, A. Al Maruf, A. Janes, and D. Taibi, “Catalog and detection techniques of microservice anti-patterns and bad smells: A tertiary study,” Journal of Systems and Software , vol. 206, p. 111829, 2023

work page 2023

[4] [4]

Towards microservice smells detection,

I. Pigazzini, F. A. Fontana, V . Lenarduzzi, and D. Taibi, “Towards microservice smells detection,” in Proceedings of the 3rd International Conference on Technical Debt , 2020, pp. 92–97

work page 2020

[5] [5]

Using microservice telemetry data for system dynamic analysis,

A. Al Maruf, A. Bakhtin, T. Cerny, and D. Taibi, “Using microservice telemetry data for system dynamic analysis,” in 2022 IEEE International Conference on Service-Oriented System Engineering (SOSE) . IEEE, 2022, pp. 29–38

work page 2022

[6] [6]

Microservice api pattern detection : Using business pro- cesses and call graphs,

A. Bakhtin, “Microservice api pattern detection : Using business pro- cesses and call graphs,” Master’s thesis, Tampere University, 2022

work page 2022

[7] [7]

Network centrality as a new perspective on microservice architecture,

A. Bakhtin, M. Esposito, V . Lenarduzzi, and D. Taibi, “Network centrality as a new perspective on microservice architecture,” in 2025 IEEE 22nd International Conference on Software Architecture (ICSA) , 2025, pp. 72–83

work page 2025

[8] [8]

Identifying hubs in directed networks,

A. Kirkley, “Identifying hubs in directed networks,” Phys. Rev. E , vol. 109, p. 034310, Mar 2024

work page 2024