Date of Award

Summer 8-2020

Document Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

Department

Computer Science

Committee Director

Michele C. Weigle

Committee Member

Michael L. Nelson

Committee Member

Sampath Jayarathna

Committee Member

Jian Wu

Committee Member

M’Hammed Abdous

Abstract

The number of public and private web archives has increased, and we implicitly trust content delivered by these archives. Fixity is checked to ensure that an archived resource has remained unaltered (i.e., fixed) since the time it was captured. Currently, end users do not have the ability to easily verify the fixity of content preserved in web archives. For instance, if a web page is archived in 1999 and replayed in 2019, how do we know that it has not been tampered with during those 20 years? In order for the users of web archives to verify that archived web resources have not been altered, they should have access to fixity information associated with these resources. However, most web archives do not allow accessing fixity information and, more importantly, even if fixity information is available, it is provided by the same archive delivering the resource, not by an independent archive or service.

In this research, we present a framework for establishing and checking the fixity on the playback of archived resources, or mementos. The framework defines an archive-aware hashing function that consists of several guidelines for generating repeatable fixity information on the playback of mementos. These guidelines are results of our 14-month study for identifying and quantifying changes in replayed mementos over time that affect generating repeatable fixity information. Changes on the playback of mementos may be caused by JavaScript, transient errors, inconsistency in the availability of mementos over time, and archive-specific resources. Changes are also caused by transformations in the content of archived resources applied by web archives to appropriately replay these resources in a user’s browser. The study also shows that only 11.55% of mementos always produce the same fixity information after each replay, while about 16.06% of mementos always produce different fixity information after each replay. The remaining 72.39% of mementos produce multiple unique fixity information. We also find that mementos may disappear when web archives move to different domains or archives.

In addition to defining multiple guidelines for generating fixity information, the framework introduces two approaches, Atomic and Block, that can be used to disseminate fixity information to web archives. The main difference between the two approaches is that, in the Atomic approach, the fixity information of each archived web page is stored in a separate file before being disseminated to several on-demand web archives, while in the Block approach, we batch together fixity information of multiple archived pages to a single binary-searchable file before being disseminated to archives. The framework defines the structure of URLs used to publish fixity information on the web and retrieve archived fixity information from web archives. Our framework does not require changes in the current web archiving infrastructure, and it is built based on well-known web archiving standards, such as the Memento protocol. The proposed framework will allow users to generate fixity information on any archived page at any time, preserve the fixity information independently from the archive delivering the archived page, and verify the fixity of the archived page at any time in the future.

DOI

10.25777/pc8d-y213

ISBN

9798678108180

ORCID

0000-0001-7648-9082

Share

COinS