HTTP Tarpit Interaction Dataset: 11-Month Analysis of Botnet Resource Exhaustion and Global DDoS Correlation

This dataset contains 251,140 interaction records captured by an asynchronous HTTP Tarpit over an 11-month period. The primary objective of the system was to exhaust the computational and session resources of attacking botnets by maintaining persistent, slow-response connections. Advanced Technical Analysis: Infrastructure: Built on a Python-based asynchronous engine (aiohttp), allowing the handling of thousands of concurrent bot sessions without system degradation. Protocol Simulation: The tarpit mimics a standard HTTP server but responds byte-by-byte with variable delays, effectively bypassing the simple timeout mechanisms of modern botnets. Global Incident Correlation: Cloudflare Hyper-DDoS (Sept 2, 2025): The sensor recorded a 5.4x increase in scanning intensity (817 events/hour) during the peak of global HTTP/2 Rapid Reset attacks. Gaming Sector Impact (Albion Online, June 2025): Captured secondary scanning waves during large-scale DDoS campaigns against MMORPG infrastructure. Secret Hunting: Detailed logs show high-frequency targeting of configuration backups (.env.prod, wp-config.php.bak) and sensitive internal endpoints. Technical Content: The provided SQLite database (tarpit_events.db) and CSV export include 23 fields covering: Full HTTP headers and User-Agent strings. Detailed GeoIP and ASN metadata. Precise connection duration and byte-count metrics. Evidence of "Secret Hunting" (targeting of .env, wp-config, and backup files). Authors & Affiliations Boiko, Viktor (ORCID: 0000-0001-5929-657X) — Scientific Supervisor & Lead Researcher. Associate Professor. Spesivtsev, Mykola (ORCID: 0009-0007-5640-5241) — Lead Software Developer & Researcher. Affiliation: Faculty of Cybersecurity and Information Technologies, National University "Odesa Law Academy", Ukraine. (ROR: https://ror.org/0282prk66)Usage and Licensing Intended for academic research in Cybersecurity, IDS Machine Learning, and Threat Intelligence. License: Creative Commons Attribution 4.0 International (CC BY 4.0).