U radu je objašnjena teorijska osnova temeljnih pojmova područja arhivistike i diplomatike te se na osnovu njih ukazuje na nedostatnost suvremenih informacijskih sustava za dugoročno očuvanje (digitalnog) arhivskog gradiva, posebno digitalno potpisanih dokumenata. U radu je pokazano da suvremeni sustavi ne uzimaju u obzir temeljne arhivističke pojmove autentičnosti i arhivske veze na način na koji to potrebno. Rad pristupa ovom problemu s pretpostavkom da ga je moguće riješiti upotrebom ulančanih zapisa te je provedeno komparativno istraživanje postojećih sustava i tehnologija. U konačnici, na temelju provedenog istraživanja, razvijen je novi model informacijskog sustava za dugotrajno očuvanje digitalno potpisanoga arhivskoga gradiva koji je u skladu s temeljnim zahtjevima arhivistike i diplomatike. The goal of this research is to explore problems related to the ability to prove the authenticity of digital archival records and especially digitally signed records in archival institutions as well as problems related to the preservation of the archival bond and to propose a solution to the identified problems. This research was motivated by previous research which has highlighted the problem of digital certificate expiry when such certificates are used to digitally sign archival records. After digital certificate expiry on which digital signatures are dependant, the ability to prove the identity of the records author is lost and the records authenticity is compromised. Contemporary digital signing certificates have a lifespan of two to three years and there is a legal requirement that archival institutions preserve their records for much longer. This discrepancy not only causes legal problems because the ability to use the record as proof of past events is compromised but also clashes with basic archival science requirements regarding archival records authenticity preservation. Existing solutions to this problem include subsequent resigning of documents and the application of digital timestamps. Digital timestamps have a significantly longer lifespan then digital signatures but suffer from the same basic problem – just like digital signatures they do not last forever. In addition to this, timestamp systems do not check or otherwise account for the identity of the author when applying the timestamp. These systems guarantee that the record has remained unchanged since the timestamp application, but the identity of the record's author is largely ignored by timestamp systems. A further problem arises form the need to maintain archival bonds, an archival science requirement which is ignored by most systems. These issues have prompted the development of a new model for long term preservation of digitally signed documents. This research began as part of the InterPARES Trust international project (2013-2019) and is concluded in this dissertation. The research began with an analysis of archival and diplomatic requirements concerning the long term preservation and the ability to prove records authenticity, including archival bond preservation. The goal of this research into archival and diplomatic literature was to establish archival science requirements for long term record preservation. Summarily said, authenticity is dependent on the ability to prove the identity of a record's author and the integrity of the record since the moment of its creation and archival bond preservation requires that the full information relating to the context in which a document was created and its network of connections to other documents are preserved. Based on these conclusions it is possible to develop a new model which accounts for archival requirements and addresses the problems related to contemporary digital signatures. Besides archival requirements, legal requirements in the Republic of Croatia were also considered. Together these requirements define the functionalities of the new model. Alongside archival requirements, cryptographic algorithms and industrial standards which govern the use of digital signatures were investigated and assessed. This includes hash algorithms, asymmetric cryptography algorithms and standards which regulate digital signature content and the application of digital signatures to records. Understanding these (cryptographic) algorithms and regulations provides clear proof of the ability to use the digital signature and related systems to prove authenticity in accordance with archival science requirements. The research also considered existing timestamp systems and analysed their advantages and shortcomings, explained why these systems do not fully fulfil archival science requirements and to which degree the systems themselves or the technology on which they are based can be used in the new model. The new model, titled TrustChain, is based on the blockchain data structure which is also used by most existing timestamp solutions. The new model has been adapted to the needs of archival institutions and considers archival science and legal requirements, primarily those connected to personal data protection. The new model enables long-term preservation of digitally signed archival records without the need for subsequent resigning or application of timestamps. In addition, the model has been adapted to long term preservation of digital certificates. The need for this feature has arisen during the conducted research and model development, mostly motivated by the confidentiality of certain archival records.
