Analysis and Extraction of Tempo-Spatial Events in an Efficient Archival CDN with Emphasis on Telegram

This paper presents an efficient archival framework for exploring and tracking cyberspace large-scale data called Tempo-Spatial Content Delivery Network (TS-CDN). Social media data streams are renewing in time and spatial dimensions. Various types of websites and social networks (i.e., channels, groups, pages, etc.) are considered spatial in cyberspace. Accurate analysis entails encompassing the bulk of data. In TS-CDN by applying the hash function on big data an efficient content delivery network is created. Using hash function rebuffs data redundancy and leads to conclude unique data archive in large-scale. This framework based on entered query allows for apparent monitoring and exploring data in tempo-spatial dimension based on TF-IDF score. Also by conformance from i18n standard, the Unicode problem has been dissolved. For evaluation of TS-CDN framework, a dataset from Telegram news channels from March 23, 2020 (1399-01-01), to September 21, 2020 (1399-06-31) on topics including Coronavirus (COVID-19), vaccine, school reopening, flood, earthquake, justice shares, petroleum, and quarantine exploited. By applying hash on Telegram dataset in the mentioned time interval, a significant reduction in media files such as 39.8% for videos (from 79.5 GB to 47.8 GB), and 10% for images (from 4 GB to 3.6 GB) occurred. TS-CDN infrastructure in a web-based approach has been presented as a service-oriented system. Experiments conducted on enormous time series data, including different spatial dimensions (i.e., Khabare Fouri, Khabarhaye Fouri, Akhbare Rouze Iran, and Akhbare Rasmi Telegram news channels), demonstrate the efficiency and applicability of the implemented TS-CDN framework.