Problems in Spatio-Temporal Modelling, Kriging, and Prediction of Computer Network Traffic
dc.contributor.author | Vaughan, Joel M. | en_US |
dc.date.accessioned | 2012-10-12T15:25:48Z | |
dc.date.available | NO_RESTRICTION | en_US |
dc.date.available | 2012-10-12T15:25:48Z | |
dc.date.issued | 2012 | en_US |
dc.date.submitted | 2012 | en_US |
dc.identifier.uri | https://hdl.handle.net/2027.42/94058 | |
dc.description.abstract | In order to maintain consistent quality of service, engineers face the task of monitoring the traffic fluctuations on the individual links making up a computer network. However, due to resource constraints and limited access, it is often not possible to directly measure the traffic on all links. This work explores a statistical framework for simultaneously modeling the traffic levels on links across an entire network and using the model to solve a variety of statistical problems, including prediction of traffic on unobserved links and the detection of statistical anomalies. We begin by examining some of the important types of network traffic data and features of the traffic. These features present interesting challenges but also provide important structure that is used throughout this work. We next develop a probabilistic spatio–temporal model for large scale computer networks that is based on physical properties of computer networks. This model simultaneously describes the traffic level on all the links of the network, and how these levels fluctuate over time. We next move on to study the so–called kriging and prediction problems, where we use observed traffic measurements on a small subsets of the links of a network to predict the traffic on other (unobserved) links in the network. We then explore an application of this prediction technique to anomaly detection. Finally, we develop an alternative model that more explicitly incorporates the dependence in traffic that arises due to certain mechanisms in the protocols that govern network behavior. We conclude by discussing the strengths and weaknesses of these two approaches. | en_US |
dc.language.iso | en_US | en_US |
dc.subject | Network Modeling | en_US |
dc.subject | Auxliary Data | en_US |
dc.title | Problems in Spatio-Temporal Modelling, Kriging, and Prediction of Computer Network Traffic | en_US |
dc.type | Thesis | en_US |
dc.description.thesisdegreename | PhD | en_US |
dc.description.thesisdegreediscipline | Statistics | en_US |
dc.description.thesisdegreegrantor | University of Michigan, Horace H. Rackham School of Graduate Studies | en_US |
dc.contributor.committeemember | Stoev, Stilian Atanasov | en_US |
dc.contributor.committeemember | Michailidis, George | en_US |
dc.contributor.committeemember | Berrocal, Veronica | en_US |
dc.contributor.committeemember | Shedden, Kerby A. | en_US |
dc.contributor.committeemember | Nair, Vijayan N. | en_US |
dc.subject.hlbsecondlevel | Statistics and Numeric Data | en_US |
dc.subject.hlbtoplevel | Science | en_US |
dc.description.bitstreamurl | http://deepblue.lib.umich.edu/bitstream/2027.42/94058/1/rsnation_1.pdf | |
dc.owningcollname | Dissertations and Theses (Ph.D. and Master's) |
Files in this item
Remediation of Harmful Language
The University of Michigan Library aims to describe library materials in a way that respects the people and communities who create, use, and are represented in our collections. Report harmful or offensive language in catalog records, finding aids, or elsewhere in our collections anonymously through our metadata feedback form. More information at Remediation of Harmful Language.
Accessibility
If you are unable to use this file in its current format, please select the Contact Us link and we can modify it to make it more accessible to you.