Authors:
Andrea Kő
;
Péter Fehér
and
Zoltán Szabó
Affiliation:
Corvinus University of Budapest, Hungary
Keyword(s):
Internal Cloud, Capacity Management, Modelling, Neural Networks, Multivariate Statistics.
Related
Ontology
Subjects/Areas/Topics:
Cloud Applications Performance and Monitoring
;
Cloud Computing
;
Cloud Computing Enabling Technology
;
Cloud Workload Profiling and Deployment Control
;
Monitoring of Services, Quality of Service, Service Level Agreements
;
Performance Development and Management
;
Platforms and Applications
Abstract:
As internal cloud, and cloud technologies widespread among companies, the responsibility of providing
reliable IT infrastructure and adequate capacities became the top priority for companies. While internal
clouds and related technologies creates the flexibility for customer, limited IT resources arise problems for
providing capacities, that has impact on IT service quality. The presented research addressing this problem,
and seeks creating models describing the relationship between IT service quality and background
infrastructure capacity usage with two distinct methodologies, in a complex cloud-like environment of a
financial institution. The research was analysed a pilot area of a widely used electronic banking service. As
multivariate statistical modelling and hypothesis testing had limited results in phase 1, but in phase 2 further
modelling opportunities were explored, a model based neural networks were developed. The research
analyses the limitations of pure statistic
al analysis in cloud-like environments, but concludes to the usability
of alternative methods.
(More)