“What the Paper Says” Critique - Cloud Services in Front-End Engineering

Muruganandam Venkatachalam, M.Sc Digital Business, University of Salford, UK

Jul 21

4 min read

Updated: Aug 7

Paper Under Review

Gu, Y. (2026). Application and optimization strategies of cloud services in front-end engineering. International Journal of Big Data Intelligent Technology, 7(1), 1–8. https://doi.org/10.38007/IJBDIT.2026.070101

Introduction

This post critiques Yiting Gu's (2026) article, “Application and Optimization Strategies of Cloud Services in Front End Engineering,” published in the International Journal of Big Data Intelligent Technology. The goal is not simply to restate what the paper says, but to evaluate how well it says it: how clearly the argument is presented, and how well the research behind it is evidenced. The paper is summarized below, then evaluated against comparable literature and direct CDN operating experience.

Summary

Gu (2026) argues cloud services benefit front-end engineering in four ways: CDNs and load balancing cut page-load times by 50–70%, citing AWS CloudFront's use across 100,000+ sites; CI/CD pipelines raise deployment frequency by over 30% and release efficiency by 40%; near-universal HTTPS, AWS IAM-style permissions, and GDPR/CCPA compliance secure data; and elastic compute, containers, and Kubernetes make systems more flexible, citing a Gartner-reported 35% scaling gain. Three equations formalize these ideas, and Gu concludes cloud services make front-end systems faster, easier to build, more secure, and more flexible.

Critical Review

The value of this paper depends less on which cloud technologies it names and more on how clearly it presents its case and how well that case is evidenced. Both are examined below, with comparable literature used to support the analysis rather than frame it.

Presentation

The paper is easy to follow and logically organized around four themes: performance, process, security, and architecture. Figure 1's flow chart of the security and privacy protection process is a genuine strength — laying out SSL/TLS, IAM, encryption, and GDPR/CCPA as one visual sequence communicates the workflow more clearly than the surrounding prose. Tables 1 and 2, however, restate the same three claims (CDN, elastic compute, load balancing) under different headings without adding new information, which makes the paper feel more thorough than its evidence actually supports.

Research Quality

The research is weaker than the presentation. Gu (2026) reads like an empirical paper but states no methodology and no explanation of how its statistics were obtained. Figures like “over 100,000 websites” on CloudFront and a Gartner-attributed 35% scaling gain appear with no traceable source. Trackers checked for this critique put Akamai's customer count anywhere from roughly 3,500 to over 40,000 companies, and Fastly's from about 600 to over 180,000 — neither reflects the vendor's own disclosed figure. The three equations use the notation of tested models but include no derivation or validation. The evidence leans almost entirely on AWS, with Cloudflare mentioned once and Fastly and Google Cloud CDN absent, and the paper never discusses trade-offs like vendor lock-in, cost growth, or cold-start latency. Independent literature is more careful on this front: Tyagi (2020) and Ccallo and Quispe-Quispe (2024) reach similar conclusions about CDN and CI/CD benefits, but ground them in defined review methods rather than unsourced figures.

A Practitioner's Lens

The clearest evidence of the research gap is what the paper leaves out operationally. In my experience running Fastly and Akamai, cache hit ratio (CHR) has run about 10% higher on Fastly, and PageSpeed testing has shown roughly a 20% page-load improvement on Fastly for the same site — neither figure appears in Gu's paper, which treats CDN acceleration as one undifferentiated 50–70% gain regardless of vendor. Architecture is a second gap: Fastly's API-first model exposes CDN, edge compute, and WAF as APIs, letting engineers define edge logic in Varnish Control Language as version-controlled code — “edge as code” rather than configuration — while other CDNs lean on managed configuration. Observability is a third: streaming logs in real time from the CDN, WAF, and edge services into platforms like Dynatrace and Splunk enables custom Real User Monitoring dashboards and Core Web Vitals tracking beyond a vendor's built-in reporting. A fourth factor is customer-side configuration: TTL settings determine cache duration — too conservative suppresses CHR, too aggressive risks stale content — and security policy works the same way, since poorly tuned WAF rules, rate limiting, and bot management introduce latency or block legitimate traffic. None of these axes rank one vendor over another; they are the kind of measurable evidence the paper should have used instead of treating “the cloud” as a single, undifferentiated category.

Conclusion

Gu (2026) is a well-organized, readable introduction to cloud services in front-end engineering, and Figure 1 in particular presents a genuinely useful workflow. Its central weakness is evidentiary rather than structural: unsourced statistics, no methodology, unvalidated formulas, and a vendor-neutral treatment that overlooks measurable differences between CDN platforms. Readers should treat its quantitative claims as illustrative rather than established, and consult primary literature — Tyagi (2020), Ccallo and Quispe-Quispe (2024) — plus direct operational experience.

References

Ccallo, M., & Quispe-Quispe, A. (2024). Adoption and adaptation of CI/CD practices in very small software development entities: A systematic literature review. arXiv. https://doi.org/10.48550/arXiv.2410.00623

Tyagi, A. (2020). Optimizing digital experiences with content delivery networks: Architectures, performance strategies, and future trends. World Journal of Advanced Research and Reviews, 7(2), 401–417. https://doi.org/10.30574/wjarr.2020.7.2.0230