
Autoscalabile ditributed anti-spam SMTP system based on kubernetes
Author(s) -
Nadja Gavrilović,
Vladimir Ćirić
Publication year - 2021
Publication title -
facta universitatis. series electronics and energetics/facta universitatis. series: electronics and energetics
Language(s) - English
Resource type - Journals
eISSN - 2217-5997
pISSN - 0353-3670
DOI - 10.2298/fuee2104525g
Subject(s) - computer science , botnet , server , computer security , system administrator , protocol (science) , computer network , the internet , world wide web , medicine , alternative medicine , pathology
Due to the increasing amount of spam email traffic, email users are in increasing danger, while email server resources are becoming overloaded. Therefore, it is necessary to protect email users, but also to prevent SMTP system overload during spam attacks. The aim of this paper is to design and implement an autoscalable distributed anti-spam SMTP system based on a Proof of work concept. The proposed solution extends SMTP protocol in order to enable the evaluation of the client?s credibility using the Proof of work algorithm. In order to prevent resource overload during spam attacks, the anti-spam SMTP system is implemented in a distributed environment, as a group of multiple anti-spam SMTP server instances. Kubernetes architecture is used for system distribution, configured with the possibility of autoscaling the number of antispam SMTP server instances depending on the system load. The implemented system is evaluated during a distributed spam attempt, simulated by a custom-made traffic generator tool. Various performance tests are given: (1) The proposed system?s impact on client?s behaviour and the overall amount of spam messages, (2) The performance of the undistributed anti-spam SMTP server during spam attack, in terms of resource load analysis (3) Autoscaling demonstration and evaluation of proposed distributed system?s performance during a spam attack. It is shown that the proposed solution has the possibility of reducing the amount of spam traffic, while processing tens of thousands of simultaneous SMTP client requests in a distributed environment.