Zhang, Heng (2019)
On the Effective Use of Data Dependency for Reliable Cloud Service Monitoring.
Technische Universität Darmstadt
Dissertation, Erstveröffentlichung
Kurzbeschreibung (Abstract)
Cloud computing is a large-scale distributed computing paradigm that aims at providing powerful computing and storage capability by dynamically sharing a pool of system resources (e.g., network bandwidth, storage space, or virtualized devices) in a multi-tenant environment. With the support of the computing technology, a plethora of cloud services have been developed for meeting the different requirements of cloud service customers (CSCs). While cloud service has many attractive advantages (e.g., rapid service deployment, reliable service availability, elastic service reconfiguration, or economic service billing), the security assurance of cloud services is a key concern for the service customers. Cloud monitoring is an essential mechanism for managing the security assurance of cloud services. Over the last few years, a large number of monitoring mechanisms have been proposed. The mechanisms are developed for monitoring varied security problems in the cloud with the common assumption that all the monitoring information is directly available. These mechanisms can achieve satisfactory monitoring performance only if the assumption can be satisfied (e.g., protecting cloud services from distributed denial of service (DDoS) attacks by checking the traffic information collected from network monitors). However, the existing mechanisms are unfortunately incapable of dealing with the security threats that are subtly crafted by malicious attackers without producing evident attack traces. Due to that the useful information related to the attacks is difficult to collect, the attacks can silently bypass the existing monitoring mechanisms and secretly undermine the victim services. As a result, to develop an effective monitoring mechanism for securing cloud services becomes a compelling demand. For motivating the issue, this thesis initially investigates a typical cloud security attack that can gradually drain system resources in a target cloud without triggering any alarms for highlighting the realistic demand of performing effective security monitoring in cloud systems. To combat the attack, a pragmatic security countermeasure is proposed for securing the cloud. To meet the demand, the thesis focuses on achieving effective security assurance management of cloud services by addressing the common shortcoming of existing monitoring mechanisms. Using the data relation (i.e., data dependency) existing in the collected monitoring data sets, the thesis demonstrates the possibility of leveraging the available information and the existing data relation to indirectly monitor cloud security problems with a novel inference-based security mechanism. Furthermore, the thesis also demonstrates the feasibility of taking advantage of data dependency to obtain the input information for running the inference mechanism by developing a practical data ascertaining technique. Finally, this thesis targets addressing potential data errors that can undermine the reliability of the proposed monitoring mechanism. Consequently, a reliability assessment mechanism is developed to select suitable data used by the proposed mechanism for generating reliable monitoring results.
| Typ des Eintrags: | Dissertation | ||||
|---|---|---|---|---|---|
| Erschienen: | 2019 | ||||
| Autor(en): | Zhang, Heng | ||||
| Art des Eintrags: | Erstveröffentlichung | ||||
| Titel: | On the Effective Use of Data Dependency for Reliable Cloud Service Monitoring | ||||
| Sprache: | Englisch | ||||
| Referenten: | Schneider, Prof. Dr. Thomas ; Suri, Prof. Dr. Neeraj | ||||
| Publikationsjahr: | 30 September 2019 | ||||
| Ort: | Darmstadt | ||||
| Datum der mündlichen Prüfung: | 18 November 2019 | ||||
| URL / URN: | https://tuprints.ulb.tu-darmstadt.de/9464 | ||||
| Kurzbeschreibung (Abstract): | Cloud computing is a large-scale distributed computing paradigm that aims at providing powerful computing and storage capability by dynamically sharing a pool of system resources (e.g., network bandwidth, storage space, or virtualized devices) in a multi-tenant environment. With the support of the computing technology, a plethora of cloud services have been developed for meeting the different requirements of cloud service customers (CSCs). While cloud service has many attractive advantages (e.g., rapid service deployment, reliable service availability, elastic service reconfiguration, or economic service billing), the security assurance of cloud services is a key concern for the service customers. Cloud monitoring is an essential mechanism for managing the security assurance of cloud services. Over the last few years, a large number of monitoring mechanisms have been proposed. The mechanisms are developed for monitoring varied security problems in the cloud with the common assumption that all the monitoring information is directly available. These mechanisms can achieve satisfactory monitoring performance only if the assumption can be satisfied (e.g., protecting cloud services from distributed denial of service (DDoS) attacks by checking the traffic information collected from network monitors). However, the existing mechanisms are unfortunately incapable of dealing with the security threats that are subtly crafted by malicious attackers without producing evident attack traces. Due to that the useful information related to the attacks is difficult to collect, the attacks can silently bypass the existing monitoring mechanisms and secretly undermine the victim services. As a result, to develop an effective monitoring mechanism for securing cloud services becomes a compelling demand. For motivating the issue, this thesis initially investigates a typical cloud security attack that can gradually drain system resources in a target cloud without triggering any alarms for highlighting the realistic demand of performing effective security monitoring in cloud systems. To combat the attack, a pragmatic security countermeasure is proposed for securing the cloud. To meet the demand, the thesis focuses on achieving effective security assurance management of cloud services by addressing the common shortcoming of existing monitoring mechanisms. Using the data relation (i.e., data dependency) existing in the collected monitoring data sets, the thesis demonstrates the possibility of leveraging the available information and the existing data relation to indirectly monitor cloud security problems with a novel inference-based security mechanism. Furthermore, the thesis also demonstrates the feasibility of taking advantage of data dependency to obtain the input information for running the inference mechanism by developing a practical data ascertaining technique. Finally, this thesis targets addressing potential data errors that can undermine the reliability of the proposed monitoring mechanism. Consequently, a reliability assessment mechanism is developed to select suitable data used by the proposed mechanism for generating reliable monitoring results. |
||||
| Alternatives oder übersetztes Abstract: |
|
||||
| URN: | urn:nbn:de:tuda-tuprints-94644 | ||||
| Sachgruppe der Dewey Dezimalklassifikatin (DDC): | 000 Allgemeines, Informatik, Informationswissenschaft > 004 Informatik | ||||
| Fachbereich(e)/-gebiet(e): | 20 Fachbereich Informatik 20 Fachbereich Informatik > Zuverlässige Eingebettete Softwaresysteme |
||||
| Hinterlegungsdatum: | 01 Dez 2019 20:55 | ||||
| Letzte Änderung: | 01 Dez 2019 20:55 | ||||
| PPN: | |||||
| Referenten: | Schneider, Prof. Dr. Thomas ; Suri, Prof. Dr. Neeraj | ||||
| Datum der mündlichen Prüfung / Verteidigung / mdl. Prüfung: | 18 November 2019 | ||||
| Export: | |||||
| Suche nach Titel in: | TUfind oder in Google | ||||
 |
Frage zum Eintrag |
Optionen (nur für Redakteure)
 |
Redaktionelle Details anzeigen |
Drucken
Impressum