Our Team Published a TLS Handshake Study: 927 Providers, Zenodo DOI
At Hitit Medya, we care about evidence-led views on digital infrastructure and web performance. A new descriptive study on TLS handshake latency from the Whspe research network—an engineering publishing group we work closely with—is now available on Zenodo. It offers a clearly bounded, single-snapshot reference for teams benchmarking hosting providers and planning SLAs.
Publication and full text
The output is descriptive (not causal) and citable via DOI: 10.5281/zenodo.19248687. The English abstract and PDF are on the Zenodo record; the full Turkish article is published on Whspe at TLS handshake live export analysis (TR).
What is being measured?
The TLS handshake is the cryptographic negotiation phase before the HTTPS application payload; because it runs before content transfer, it directly contributes to first-connection delay. The reported numbers are measured milliseconds from the supplied live export, not a theoretical model.
Headline findings (summary)
- Scope: 927 hosting providers and 9,210 TLS observations (live export dated 27 March 2026).
- Robust centre: A 10%–10% symmetric trimmed mean yields a global summary of 57.31 ms, versus a raw mean of 81.16 ms.
- Cross-check: A second 10%–10% trim on provider-level trimmed means gives 57.52 ms, consistent with the global trimmed estimate.
- Distribution: P10 = 10 ms, P50 = 51 ms, P90 = 192 ms, IQR = 66 ms.
| Metric | Value |
|---|---|
| Raw mean | 81.16 ms |
| 10%–10% trimmed mean | 57.31 ms |
| Median (P50) | 51 ms |
| P90 | 192 ms |
Why a trimmed mean?
When TLS latency is heavy-tailed, the raw mean can be pulled upward by extreme values. The study positions the trimmed mean alongside the median and percentile bands as the primary robust summary, while still reporting the raw mean for transparency.
Authors and license
The Zenodo record lists Tahir Dinç (data manager) and Celal Dinç (project member). The work is released under Creative Commons Attribution 4.0 International (CC BY 4.0).
This post is an announcement summarizing a single-snapshot descriptive study that does not claim causality. For methodology, limitations, and regional detail, please read the full paper.
