Long-Term Preservation of Digital Signat
Long-Term Preservation of Digital Signat
Long-Term Preservation of Digital Signat
Hrvoje STANI*
LONG-TERM PRESERVATION OF DIGITAL SIGNATURES
Abstract:
Digitally born documents are increasingly being signed by digital signatures. Once becoming records,
they need to be preserved in some cases for several years while in other cases permanently. In
essence, what should be preserved besides digital records themselves, is their trustworthiness. This
may be a problem because digital records might be converted to new formats, migrated to new media,
emulated or virtualised in new environments due to the technological obsolescence. Authenticity, in
particular, relies on the possibility to check the validity of the digital signature. However, digital
signatures expire after a certain period. The author investigates the possibilities of long-term
preservation of digitally signed records in the ever-changing IT environment.
Key words:
digital records, digital signatures, digital certificates, long-term preservation, archival science
Izvleek:
Dolgorono varovanje digitalnih podpisov
Izvorne digitalne dokumente vedno pogosteje podpisujemo z digitalnimi podpisi. Ko tako postanejo
del zapisa, jih je potrebno ohraniti obiajno nekaj let, v nekaterih primerih pa tudi trajno. Poleg
samega digitalega zapisa je nujno ohraniti tudi njegovo verodostojnost. To lahko predstavlja problem,
saj so lahko digitalni zapisi pretvorjeni v druge oblike, preneseni na nove medije ali virtualizirani v
novih okoljih zaradi zastaranja tehnologije. Avtentinost dokumenta e posebej sloni na monosti
preverbeveljavnosti digitalnega podpisa. Le-ta pa po doloenem obdobju preneha. Avtor raziskuje
monosti dolgoronega varovanja digitalno podpisanih zapisov v vselej spreminjajoem se IT-okolju.
Kljune besede:
digitalni zapisi, digitalni podpisi, digitalna potrdila, dolgorona hramba, arhivska znanost
INTRODUCTION
Hrvoje Stani, Ph. D., associate professor, Department of Information and Communication Sciences, Faculty
of Humanities and Social Sciences, Ivana Luia 3, Zagreb, Croatia, contact: hstancic@ffzg.hr.
481
fundamentally changes once the digital signatures are introduced. First of all, as it
will be presented later on, every digital signature from the same author is different
from one another. Then, there may be infinite number of originals since making a
digital copy of a digital original may maintain every characteristic of the original
and may function as an original. Therefore, we cannot apply the concept of original
from the analogue to the digital world. We have to change the paradigm regarding
the originality since the authenticity of the digital signature cannot be verified
without a third party, as it will also be argued later on. Considering all this, it is
important that archivists understand the mechanisms behind digital signatures in
order to be able to verify authenticity, integrity and non-repudiation of a message,
document or a record and to be able to reach an informative decision in the context
of (long-term) digital preservation.
2
DIGITAL SIGNATURES
482
The author can also calculate the so-called document digest by applying the
hash algorithm to the document (e.g. using MD5 method). The result is a 128-bit
number,1 which is unique for every document. If a document changes, its digest
would change too. Friedl (2005) argues that the creation of the document digest is
collision resistant, i.e. that there are no two different documents for which the
same digest would be calculated. Also, it is impossible to reverse the operation and
get the contents of the document by knowing the document digest.
For example, hash value calculated from the abstract of this paper using MD5
method is: b998ff7b2695ad3d22233ed226fdfbe3. If only one letter is changed,
e.g. if the initial word Digitally is changed to digitally, the MD5 method would
generate a substantially different hash value, e.g.: 72fe719ffbf0cf7d5e8784
dda3e87bbc, for the whole abstract.
When the author wants to digitally sign a document, he applies his private key
to the document digest and sends it to the recipient. The recipient recalculates the
document digest and by applying the author's public key to the document digest,
verifies if the author has signed the document. This is a faster procedure with the
same effect since the signature is either confirmed or refuted by checking against
the document digest and not the whole document, which might be of considerable
length.
The only thing we are still not sure is who is the real person standing behind
the digital signature. There are numerous web services offering the creation of a
combination of public and private keys and anyone can claim to be anyone else if
(s)he wants it. Therefore, we need digital certificates.
3
DIGITAL CERTIFICATES
If we want to be sure that the author behind a digital signature is indeed that
person and not someone else impersonating the author, we need a trusted third
party called certification authority. Boettcher and Powell (2002) compare digital
certificates to virtual ID cards issued by a trusted authority and explain that a PKI
includes organizations called certification authorities (CAs) that issue, manage, and
revoke digital certificates. (...) A CA might create a separate registration authority
(RA) to handle the task of identifying individuals who apply for certificates. Google
Chrome help section explains that public key certificate, usually just called a
certificate, is a digitally signed statement that binds the value of a public key to
the identity of the person, device, or service that holds the corresponding private
key. Microsoft (2005) further clarifies that because the certificate matches a
public key to a particular individual, and that certificate's authenticity is
guaranteed by the issuer, the digital certificate provides a solution to the problem
of how to find a user's public key and know that it is valid. These problems are
solved by a user obtaining another user's public key from the digital certificate. The
user knows it is valid because a trusted certification authority has issued the
certificate. In addition, digital certificates rely on public key cryptography for their
own authentication. When a digital certificate is issued, the issuing certification
authority signs the certificate with its own private key. To validate the authenticity
of a digital certificate, a user can obtain that certification authority's public key
MD5 (Message Digest) produces 128-bit hash value, SHA-1 (Secure Hash Algorithm) produces 160-bit and the
more advanced SHA-2 produces 224, 256, 384 or 512-bit hash value.
483
and use it against the certificate to determine if it was signed by the certification
authority.
Microsoft Windows 7 help section explains that the path validation involves
processing public key certificates and their issuer certificates in a hierarchical
fashion until the certification path terminates at a trusted, self-signed certificate.
Typically, this is a root CA certificate. If there is a problem with one of the
certificates in the path, or if it cannot find a certificate, the certification path is
considered a non-trusted certification path. A typical certification path includes a
root certificate and one or more intermediate certificates.
484
the subject's identifier information, such as the name and e-mail address,
the validity period (the length of time that the certificate is considered
valid),
the digital signature of the issuer, which attests to the validity of the
binding between the subject's public key and the subject's identifier
information.
However, Brzica, Herceg and Stani (2013, p. 149) point out that the
Directive 1999/93/EC allows issuing of the so called qualified certificate which is
based on the RFC 3039 standard and implements the concept of non-repudiation
and enumerate 10 elements that a qualified certificate must include.
Digital certificates may be valid or may, for a particular reason, be revoked.
The certificate revocation list (CRL), according to TechTarget (2007), is a list of
subscribers paired with digital certificate status. The list enumerates revoked
certificates along with the reason(s) for revocation. The dates of certificate issue,
and the entities that issued them, are also included. (...) The main limitation of
CRL is the fact that updates must be frequently downloaded to keep the list current.
Online Certificate Status Protocol (OCSP) overcomes this limitation by checking
certificate status in real time. In that context Brzica, Herceg and Stani (2013,
pp. 150151) explain that non-repudiation is a characteristic of a record that
prevents any signatory to deny the action taken or the content of a record. In the
Croatian legislation non-repudiation is associated with the advanced digital
signature which is based upon qualified certificate. For a record to achieve and
preserve characteristic of non-repudiation it is necessary to ensure:
1. digital identity of signatories,
2. real-time revocation of digital signature rights,
3. time-stamping of digital signatures after checking the list of revoked
certificates, which ensures the validity of electronic signature at the time
of signing, and
4. secure long-term preservation.
485
LONG-TERM PRESERVATION
In this paper the terms digital archive and digital/preservation repository will be used with the same
meaning.
486
The first and second usages are common today as digital signatures are used
in the transmission of business documents and other data. Typically, validation
takes place shortly after signing and there is no need to preserve the signature itself
over time. In the first case the repository may record the act of validation as an
event, and save related information needed to demonstrate provenance in the
event detail. In the second case the repository might also record the signing as an
event but the use of the signature is the responsibility of the receiver. Only in the
third case, where digital signatures are used by the repository as a tool to confirm
the authenticity of its stored digital objects over time, must the signature itself and
the information needed to validate the signature be preserved.
According to Blanchette (2006, p. 14), from the point of view of archives,
there are three possible solutions:
1. Preserve the digital signatures: This solution supposes the deployment of
considerable means to preserve the necessary mechanisms for validating
the signatures, and does not address the need to simultaneously preserve
the intelligibility of documents;
2. Eliminate the signatures: This option requires the least adaptation from
archival institution, but impoverishes the description of the document, as
it eliminates the signature as one technical element used to ensure the
authenticity of the documents;
3. Record the trace of the signatures as metadata: This solution requires
little technical means, and records both the existence of the signature and
the result of its verification. However, digital signatures lose their special
status as the primary form of evidence from which to infer the authenticity
of the document.
In order to preserve digital signatures along with the records, the archives
should have the possibility to validate the signature at any given moment in the
future. Due to the fact that digital signatures and associated certificates are valid
only for a certain period of time and that the certificates might be revoked, as it
was discussed before, this option seems rather unlikely to function at the long run
unless certain preconditions are being met. Dumortier and Van den Eynde (2002)
argue that the only effective solution (in their view) for the problem of signature
durability, is the archival of the original binary representation of the document.
This solution was proposed by the European Electronic Signature Standardization
Initiative (EESSI) in the study report Trusted Archival Services (TAS). A TAS must
guarantee that it will still be possible to validate archived document years after the
initial archival date, even if the applications that have been used at signature
creation time are no longer in use. In other words, the TAS should maintain a set of
applications (viewers as well as signature validation applications) together with the
corresponding platforms (hardware, operating systems) or at least an emulator of
such applications and/or environment in order to guarantee that the signature of
the document can still be validated years later. It is clear that this option would
require a lot of technical skills and expertise from the archives, not to mention the
financial implications.
The second option is technically the least challenging in the context of longterm preservation but it is actually not an option for archiving of the records that
need to be preserved as authentic.
487
CONCLUSION
records are kept in a trusted archival environment and that no either unauthorised
or authorised (preservation) procedure has influenced the aspect of
trustworthiness3 of the records.
Nevertheless, there are situations where the records are being kept in the live
systems, or simply stored for quite some time before submitted for archiving. It
might be that the digital signatures and the associated certificates are already
expired at that moment and, as suggested in the earlier discussion, the archives
could either ingest them without the authenticity information added to the
metadata or refuse to ingest them. This could pose a problem if the law requires
the records to be preserved and if the archives rejects them due to the non-validity
of the digital signatures. Therefore, it would be a challenge to further investigate
the possibilities of prolonging the period of validity of digital signatures.
6
FUTURE RESEARCH
The author will, along with other partners and as part of the recently initiated
research topic at the InterPARES Trust project4, investigate the possibilities of
revalidation of the expired digital signatures, periodical re-signing of digital
records, addition of timestamps, injection of additional (timestamped) proof of
existence, etc.
RESOURCES
Blanchette, J.-F. (2006). The digital signature dilemma. Annales des Tlcommunications, 61
(78), pp. 908923. Accessed 6 January 2016 from: http://polaris.gseis.ucla.edu/
blanchette/papers/annals.pdf.
Boettcher, J. V. and Powell, A. (2002). Digital Certificates. CREN. Accessed 5 January 2016
from: http://www.cren.net/crenca/docs/syllabus.pdf.
Data Dictionary for Preservation Metadata: PREMIS version 3.0. Accessed 3 January 2016
from: http://www.loc.gov/standards/premis/v3/premis-3-0-final.pdf.
Dumortier, J. and Van den Eynde, S. (2002). Electronic Signatures and Trusted Archival
Services. Proceedings of the DLM Forum: Access and preservation of electronic information:
best practices and solutions Barcelona 6-8 May 2002 (pp. 520-524). Luxembourg: Office for
Official Publications of the European Communities. Accessed 15 March 2015 from:
http://www.expertisecentrumdavid.be/davidproject/teksten/DAVIDbijdragen/Tas.pdf.
Foscarini, F. (2008). Cultures of Trust: Legal, Technical and Archival Perspectives on the Use
of Digital Signature Technologies. Hegering, H.-G. et al (Eds). INFORMATIK 2008
Beherrschbare Systeme dank Informatik Band 1, pp. 3747. Bonn: Gesellschaft fr
Informatik. Accessed 7 January 2016 from: http://subs.emis.de/LNI/Proceedings/
Proceedings133/gi-proc-133-007.pdf.
The concept of records' trustworthiness refers to authenticity, reliability, accuracy, integrity and usability of
records.
489
Groven, A.-K. (2010). Trust Strategies in Longterm Management and Preservation of Digital
Records. A Deliverable to the LongRec Research Project. Norsk Regnesentral. Accessed 18
December 2015 from: http://publications.nr.no/directdownload/publications.nr.no/5457/
Groven_-_Trust_Strategies_in_Longterm_Management_and_Preser.pdf.
Stani, H., Rajh, H. and Brzica, H. (2015). Archival Cloud Services: Portability, Continuity,
and Sustainability Aspects of Long-term Preservation of Electronically Signed Records.
Canadian Journal of Information and Library Science, 39 (2), pp. 210227. Toronto: University
of Toronto Press.
TechTarget (2007). Certificate Revocation List (CRL) definition. Accessed 5 January 2016
from: http://searchsecurity.techtarget.com/definition/Certificate-Revocation-List.
POVZETEK
Hrvoje STANI*
DOLGORONO VAROVANJE DIGITALNIH PODPISOV
Digitalna komunikacija je danes postala standarden nain komuniciranja.
Prisotnost podpisa in drugih oblik avtentinosti, kot so npr. igi, je razlikovala
originale od kopij. Nasprotno od analognih podpisov pa se vsi digitalni podpisi istega
avtorja med sabo razlikujejo. Prav tako lahko obstaja neskonno tevilo digitalnih
originalov, saj digitalna kopija digitalnega originala ohrani vse znailnosti originala
in lahko kot taka tudi nastopa.
Koncept digitalnih podpisov sloni na konceptu infrastrukture javnih kljuev
(Public Key Infrastructure - PKI). Avtor generira dva kljua zasebnega in javnega.
Zasebnega ima vedno pri sebi, javni pa je dostopen javnosti. Obstajata tudi dva tipa
digitalnih podpisov osnovni in napredni, izraena pa sta lahko v mnogih formatih
(XMLDSig, XAdES, CAdES, PAdES). e elimo biti prepriani, da za digitalnim
podpisom stoji pravi avtor in ne nekdo, ki se zanj izdaja, potrebujemo e zaupanja
vredno tretjo osebo, tj. overitelja, ki izdaja, upravlja in preklicuje digitalna
potrdila. Potrdilo je digitalno podpisana izjava, ki povezuje vrednost javnega kljua
z identiteto osebe, naprave ali storitve, ki je lastnik pripadajoega zasebnega
kljua. Kvalificirano potrdilo prinaa tudi koncept nezatajljivosti. Digitalna potrdila
so lahko veljavna ali preklicana in se tako znajdejo na seznamu preklicanih potrdil.
Za arhivsko znanost in prakso je poznavanje problema pomembno, saj se
digitalno podpisani zapisi e zajemajo v digitalne arhive, to pa lahko pomeni
precejen izziv v kontekstu dolgoronega varovanja. Arhivist se mora odloiti, kaj
narediti z njimi, in hkrati razmiljati o tem, da imajo digitalni podpisi z arhivskega
vidika kratek rok veljavnosti in so lahko kadarkoli preklicani. Vse to vnaa dvomv
arhivski proces in ovira monost preverjanja avtentinosti, celovitosti in
nezatajljivosti zapisa po doloenem asu. S stalia arhivov obstajajo tri mone
reitve: (1) ohraniti digitalne podpise, (2) izloiti podpise ali (3) zapisati sled
podpisa kot metapodatek. Za dolgorono varovanje avtentinih digitalno podpisanih
*
Doc. dr. Hrvoje Stani, Oddelek za informacijske in komunikacijske znanosti, Filozofska fakulteta, Ivana
Luia 3, Zagreb, Hrvaka, kontakt: hstancic@ffzg.hr.
490
491