Markovian chain-based network availability analysis considering security issues

UDC 303.733.33

Markovian chain-based network availability analysis considering security issues
Alaa Mohammed Abdul-Hadi, PhD-student
National Aerospace University named after N.E. Zhukovsky “KhAI”

Most of the operational computer networks, systems, and service oriented architectures (SOA) need to provide continuous services to their subscribers. Otherwise, service unavailability may lead to severe consequences for the provider and its reputation, in addition to, the national economy, security and safety, if the effected IT services are time-critical services, upon which, the critical information infrastructure, rely on.

Thus, availability need to be evaluated during the design, implementation, operation and development to minimize risk of service downtime.
There are many existing quantitative, analytical model-based techniques for availability evaluation. These techniques do not consider or not enough deeply consider (in point of view quantitative assessment) security issues, such as, attacks and vulnerabilities which represent a serious threat to networks and their provided services causing an additional source of outages.
This leads to inaccurate results obtained from solution of these methods due to the neglection of the security issues during all phases of the evaluation process. Thus, to obtain a more acceptable and realistic results a new model is needed that considers the mentioned security issues.
In this paper, we review existing quantitative model-based evaluation techniques for availability and security. Then, a new approach based on Maikov model to evaluate network availability considering different types of attacks and vulnerabilities will be developed. Finally, we apply our approach to different networks and web-service architectures in order to satisfy the obtained results.
*Научный руководитель D.Sc., professor, V Kharchenko