Validate digital sign template
(upbeat music) Hello , my name is
Bernd Wild from Intarsys and I will take you on a tour through the world of signing
and validating PDF documents. I would like to give
you a better impression
of what happens when the validating a
digitally signed PDF document. And we will see that validating
a signature in a PDF can be a real challenge. And that validation as a whole
is often more complex than actual signing. Well, what is validation? Validation of a
digital signature, regardless of the
document formart always consists
of two components, proof of the identity
of the signer, and proof of the integrity
of the signed content. A digital signature
is only valid if both checks are successful. The scope of verification of
a digital signature, goes far beyond that of a hand
written signature. The case of paper, it is only possible to
check with great efforts and with the help of experts, whether any changes have been
made to the written content after signature has been fixed. Just think of multi-page
contract with a signature on the last page and the
identity of the signer that can actually
not be verified by the recipient at all. He's usually dependent on the
believing that the printed name of the sign-on matches
the handwritten signature. digital signatures therefore offer a completely different
level of security and trust. The case of digital signatures, the identity verification of
a signature is based on the verification of the certificates
and the certificate chains. And is described in
detail in standards such as carters and charters. This is therefore not the
focus of the presentation. We will rather concentrate
on the second part, the proof of content integrity. whatever one wants to
achieve when opening, or validating a digitally
signed PDF document, is the green check mark. This is the most
important thing. Here I have shown this green
check mark on two different PDF viewers, left in Adobe reader, and right in sign live
CC our own PDF viewer. We notice two points. First, validation takes place
automatically without any further action when
the PDF document is opened in the viewer. That means the user does not
have to initiate an explicit validation process. Of course he can. And secondly, we can
already see various sub-items in the check result. Which illustrate that the check
process itself consists of several steps. Only when each of these
step provides a green hold. The overall result
is also positive. Unfortunately, it
is quite possible that some steps do not
have clear results. Then the green check
mark quickly turns
into an exclamation mark, that requires a closer
examination of the validation result. What you've never across, which indicates a
negative validation. The assessment, what went wrong is
then up to the user. And from personal experience, this is often the point at
which the hotline of the manufacturer of the viewer
or signature application is contacted by the
user to determine the
cause more precisely. Because to be honest, the interpretation of the
validation report of a digital signature is not
exactly user-friendly. The underlying processes
are too complex for that. So the best thing to do is
to make sure everything runs smoothly and ends up
in a green check mark. PDF as a document format
supports several types of digital signatures, which differ purely in
their validation behavior. With regard to the
changeability of content after a signature has been applied. technically and
cryptographically, both signature
types are identical. On the one hand, we have the most common and
widely used signature type, the so-called
approval signature. You see it here on the
right side of the table. This is the standard
signature form in PDF. It can be applied to a PDF
as many times as desired. One after the other to obtain
so-called serial signatures. An approval signature,
always refers to the entire PDF document. At the time of the
signature to be precise . focusing only on certain
objects in the PDF or often asked on
individual pages, is not possible with
approval signatures. This special form was originally
possible directly after the introduction of digital
signatures in PDF with PDF version 1.3, but was then removed by Adobe
with PDF 1.6 and has been prohibited since then. Approval signatures
allow mark and anotations to be applied to
the document after signing, without affecting the
validity of the signature. This means that commands and
graphic annotations can be applied. Furthermore, the feeling of
form fields is also allowed. However, changing the shape
or form fields is not allowed. As mentioned, this is the standard
case without any further specifications. The certification signature on
the other hand, is a limited form of the approval signature
in terms of modifications. Here, the scope of modifications that is still permitted after signature can be
specifically limited. Mind you restricted,
not extended. It's not possible to
extend the possibility the author signature. An approval signature cannot
create additional modification possibilities. And as a special feature, a
certification signature must always be the first
signature in a PDF document. Hence the alternative
name of a signature. It is not allowed to add a
certification signature after your approval signature
or multiple certification signatures. This will inevitably lead to
a validation error, as long as the PDF viewer and validator
is compliant with the PDF standard. Here is a brief list of the
most important standards that play a role in the creation
validation of PDF signatures. Of course, there are many more as these
standards are based on other specifications as well, but for now these are
the essential documents. Besides the three-part ETSI, and SA standard 31942, which describes the
so-called PAdES standard, which stands for PDF based
advanced electronic signatures, the ISO 32,000-1 and
-2, also include the way signatures are integrated
into the PDF structure. Only very privy to the
process of signing a PDF, which is shown here
in a simplified form. The PDF document is to
be teacherly signed, the certificate of the signer and his public and
private key are required. Usually for a high quality
advanced or even qualified signature. The private key is stored
on a highly secure hardware, such as a smart card or an
HSM hardware security module, such as those for use by
remote signature services. The first step is to split
the existing PDF at a suitable location, insert an empty block,
and then save the document again.The new PDF document
now consists of three parts. The empty block part two, is a
space that is taken up by the signature container later on. Here called CMS container. After the signature
has been created , the coordinates part one and part three from the content of the
so-called byte range, which is written at the
beginning of part three, then a hash value
is calculated over the range of part
one and part three. With the help of
the private key, this hash value is then
encrypted and forms the signed hash value, the
actual signature value. This hash value is packed into
the CMS container together with certificate information, validity information
for the certificates according to the
carter standard. If you open a signed
PDF in a text editor, or in a cross-browser, you can easily find out
where the CMS container was inserted. A signature in the
PDF is implemented
by an acro form field of type signature or/sick As you see here in this slide. The value of the dictionary
element contents is the base 64 encoded value of
the CMS container. The byte range directly
following the contents is also nicely displayed. In this case, part one
of the PDF starts at byte zero and has a length
of 84,102 bytes. While part two starts
at byte position 160,872 and has alength of 775 bytes. If you take a closer look at
the CMS container and decode it with an appropriate tool, you will recognize the
hierarchical ASN.1 structure, which includes all signature
of relevant information. This means that almost all
signatures and certificate information, including the encrypted hash
value is located within this container. Therefore, the structure of the container
is also independent of the actual document format and
would look exactly the same for assigned film or image. PAdES, is essentially only the transfer or the Carter
standard to the PDF case. Here. Again, the structure of the byte range. Only the range specified here
is covered by the signature, and can therefore easily
be checked for changes. Every change within the range defined find by the byte range, always an inevitably leads to
a validation error and thus, to the breaking of the signature
and this without exception, remember it is not possible to
sign only individual objects within a PDF document. This possibility is no longer
available since PDF 1.6. If you still want to allow
modifications in the PDF, After a signature
has been applied, you have to resort
to another trick, which I will
discuss in a moment. What exactly happens when
validating such a signature? Here are the steps that are
taken to verify the identity of a signature. Of course, this can only be
understood as an example, since the steps can vary, depending on the selected
signature profile. The result of the check provides
information as to whether the said signer has
a certificate issued
by an authority recognized as trustworthy, and whether this certificate
was valid at the time of the signature. Depending on the specifications, the root certificates in
the so-called trust list are considered trustworthy. In case of qualified
electronic signatures, according to the eIDAS
regulation, this would mean, for example, the national eIDAS trusted lists and the EU's list of
eIDAS trusted lists. In the PDF world, the Adobe Approved Trust List AATL ,is also regarded
as an anchor of trust for verifying trustworthiness. And here are the steps to
determine the integrity of the content. The crucial step here is
the last one the list. Where both hash
values are compared, that means the hash value of
part one and three immediately before the signature, and the count hash
value of this areas at the time of validation. If both equal, then
the integrity check
has been successful. If the content of the PDF cannot
be modified after signing, without breaking the signature, how can modifications be
made without affecting the validation algorithm? Of course, this is also includes applying
other signatures in the case of serial signatures.
Since as we have learned, applying a signature
is done by creating a corresponding acro form field, with it's substructures
in the PDF. The trick, is to append new
revision to the existing PDF. This way, the already signed
area is not touched or changed. So the use byte range
can still be used to calculate the hash value. If a new signature is appended, the same mechanism is performed
within the new revision as described here earlier. That means ,get formation
and determination of the new byte range. Which now completely includes
the previous revision, and the back half
of the new version. Then the second signature
container is inserted into revision one. This process can be repeated
as often as desired. This allows as many
signatures to be applied one after the other. It means serially 'S' is useful
for the business workflow. Up to now we have always
done the process of splitting into two parts when applying the
first signature
directly in the body of the PDF. That means without
ending a new revision, The usual aim is another round. To be precise however, a new revision is already
added with the first signature and the signature container
becomes part of this revision. This is shown again
in this slide. That means you can take
a look for example, at our application Sign Live! CC And then you will notice
that if a single signature is applied without any
further changes, only one additional revision
is added to the PDF. In this picture, this
would be revision one. The exciting part now comes
when the permitted modification is to be made to
the CND document. The case shown, a command
is added to the signed PDF. As already shown for
serial signatures, the only way to implement this
in the signature compatible way, is to append a new
revision, revision two in this case ,to the PDF
and define the command as an annotation within its body. This means that in the
body of revision two, all necessarily PDF objects on the cost level
must be defined. As well as the changes
to the objects from revision CRO, such as a page objects, so that the
command can be visualized in the viewer at all. The corresponding sections
from the ISO 32,000 specifications that govern the
handling of annotations and signatures, and that compliant validation
are presented here. It is clearly stated that
in case of a subsequent modification, and this plate state
does not correspond to the science state. If however, the
annotations are used, according to the description in chapter 12.5 of
the ISO specification, this does not lead to a
breaking of the signature. On the contrary,
annotation is recommended as a means of choice for
implementing document workflows. What exactly happens
when assigned and
subsequently modified PDF is validated can be
summarized under the term modification analyzers. Unfortunately, exactly this process
is not clearly and unambiguously despecified. This means that there is hardly
any specification material for someone who wants to
implement such a behavior. The benchmark here are the
implementations of Adobe reader and Acrobat. For everyone else,
this is more or less a realistic process, which may lead to more or less
discussions with customers, and users are even too
generous and tolerant interpretations, which in turn are more
vulnerable to malicious attacks. The general goal is to
avoid the green hook, for a compliant
modified signed PDF. Let's take a closer look at
the modification analyzers. We start with a document
that contains any remissions. For each revision, we check
whether a signature is included. If so, it's really
duty is checked by the respective byte range. Afterwards, it is
determined whether there is a subsequent
revision that contains mark up annotations
or another signature. If mark up annotations
are present, they're checked for the effect
on objects from the previous revision. This means for example, that the previous page object, is only changed to
include the new annotation Additional non-permitted
objects may not be inserted. This is followed by a check
of the cross-reference table. The last step when it is checked whether certificate signature
is present. And if so, whether the modification
restrictions
possibly stored there are violated. At the end of this
analyzers process, either a green check mark is
displayed to indicate that the signatures are valid, and only
permitted modifications have been made, or an error
messages displayed indicating that the requirements or restriction or the
approval or certification signatures have been violated. If we now add several
annotations to this document, the validation result
itself remains the same. However, it is noted here that several
changes have taken place after the signature. However, these changes have not
affected the validity of the signature, which is why the green check
mark in front of the signature applies still in this picture. The present case several
annotations were applied. For example, the rotor test was marked yellow Furthermore ,a form field was added and red borders
were drawn around individual objects on the page. This brings us to another very
important use case of post assigning modifications, the so-called
signature augmentation. at HCN standards provide for
certain signature profiles. In which all certificate
chain informations necessary for
identity verification, that means validity and
revocation formation like OCSP responses or CRL information,
can be stored together with a signature so that
it is necessary. A complete validation of the
signature can be performed without any online access to
the trust service providers involved in the certificates. In this case, one can speak
of a self-contained signature, which apart from the durability
of the crypto graphic algorithms used,
its very well-suited for long-term archiving. Since no further
external resources or
online instances need to be consulted for
future validation. For PDF signatures, this
extended validation information can be stored either in
the CMS container itself, and thus transparent for
the pre-PDF document, or in specially
created dictionaries. These include the DSS
or document security
store dictionary, and the VRI, the validation
related information dictionary. These two dictionaries were
explicitly specified for the first time in the
standard EN 319142-1 as part of PAdES ,and
are not in ISO 32,000-1 but firstly appear
in ISO 32,000-2 Excalty this possibility of allowed post signing
modifications is used by a class of
attacks on signed PDFs that became known as
shadow attack this year. Still for us to manipulations of revisions have
signed PDFs, that use was either incorrect or
corrupt revision structures or annotations, that visually
overlay the original signed content in such a way
that at the first glance, the originally signed content
appears to have been changed. Similar to the
previously published a
text on the signature structure, also known under the German
universal signatory forgery. The good nature of many
PDF viewers and validators is exploited to
interpret corrupt PDF
structures in order to accommodate the user
as much as possible. Together with a lake
of exact specifications on how to perform
the modification analyzes, positive validation
results can result despite malicious modifications. Of course, attack scenarios
presented@pdfinsecurity.org
require very specific
constellations in the PDF, which are not necessarily
found in the market. Rather they are theoretically
possible methods of compromising assigned PDF
with respect to validation. As long as the PDF tools and validators strictly
adhere to the specification, there no significant risk. Even if in some implementations, the user experience is more
important than the correctness of the format. Nevertheless, we must be grateful to
the authors of the study. For specifically pointing out
possible improvements with regard to trust in
digital signatures in PDF. They describe problems with
the modification analyzes and flat to the fact that they are
now also taken up by ETSI as a standardization body for
electronic signatures and their application to PDF. Several work packages
are currently working now, on the
specifications of
these requirements and their definition
in a supplement to
the known standards. The result of this work is
expected to be ready in 2021, so that we can assume, that
the validation of signed PDFs will then become somewhat
more transparent and no longer pose the challenge
as it does today, we at Intesis are manufacturers
of software for the electronic signature
of documents and data. This includes the entire range from biometric to qualified
electronic signatures. The core product here is
Sign Live! CC for use at the workplace and Sign Live! cloud suite gears for the web-based signature. Intesis supports a wide
range of signature devices such as smart cards,
crypto tokens, and HSMs from various
manufacturers. Sign Live! is used
across all industries in various Europeancountries, in cooperation with various
trust service providers. We are also actively involved
in various signature related standardization committees, such as ETSI and
the cloud signature
consortium, of which we are also a founding member. As one of the first members
of the PDF association, we are represented on the
board of directors of the association and focus on the
work around the digital PDF signature. Our slogan for Sign Live!
is 'To be better than ink.' That means to improve user
experience when making the transition from handwritten
and signatures ,or paper-based signatures, to
digital signatures, Just a few facts
about the company. And we have about around 17
long time employees and more than 1100 customers
from various industries and public administration. We work together with
more than 40 partners. And with our software, the customers produce more
than 5 million signatures per month, and they are doing also more
than 5 million validations per month. In total, we have relationships to around 18 European trust
service providers, and we are supporting actively
around five to six with our APIs directly,
others with a short time of integration, which
is around two to three days. This again is a short overview
about our portfolio and besides signature software, we also involved in PDF
and PDF aid development, especially for the usage within
electronic invoicing and the the standards so-called
scaffold in Germany or Factor X now in Europe,
France, and Germany. We also offer consulting
in these domains, That means if you need some
consulting for developing an appropriate signature solution for integration of an
signature solution or, and PDF or PDFA solution, you can also come to us
and we will help you. And we support our customers
in a holistic way with extensive know-how and of
corresponding component, competence and offer them
solutions from one source. So they can ask us, and we are looking for all
the different interfaces, and we will have them to
integrate their software and to help them with the problems. From the trust services
as defined by eIDAS, we focus on the first
categories of trust service, electronic signatures,
remote signatures, as well as the electronic
seal or remote seal, that are trying timestamp. And last one, the
preservation services. We don't offer a products
for electronic trusted email services or website
authentication. This is outside of our scope. They are in this
four categories. We support secure signing,
secure validating, and secure preserving. That means if you need to
preserve your documents in a cryptographical way for
more than 10 or 15 years, that means after the expiry of the actual cryptographical
algorithms, we have the solution
for this to do it, So this, this so-called
IQ soft product, which gives you history based renewal of already
assigned documents. Some of our customers
are shown here, So we work with a lot of big,
medium scale cam companies, and we have them as customers, Some of them also, as, as partners, for example,
we are working closely with, with Swisscom in
Switzerland, as well as Bundesdruckerei
with in Germany, because they are the main
trust service providers here. We also support from two
systems or Tailor SEC, they are trust
service provisions, and then you trust
services coming up for, from Bank for luck and
also from a Me Design. Signing with Sign
Live!,our corporate act as I said, you can do this. And on the local base with a
smart card or a crypto token, as well as on the remote base, that means then you have to
use a server or you go and use our client and the client talks directly
to your remote signature service. And there we can
offer all common more signature
services in Europe, as, as long as they are
published via the CSE API. Beside this, we can give you also
solutions for local or remote times stamping and means if
you want to set up your own timestamp server for
advanced the timestamps or even for qualified ones, we can deliver you the complete
software stack to do this, and to have an app as a
very reliable and performant installation for times
stamping your documents. Sign as you like, as with that, you can choose
among these offers, what you want to implement,
you can use, as I said, time stamping in a
local or remote manner, you can have the
personal signature based on software assisted
certificates on smart cards, or even on the remote
signing service, and the same applies to seals. And you can have this also in, in the quality
signature qualities,
qualified or advanced. And if you like, you can also include your
handwritten signature taken by a pressure sensitive PET, and in your digital signature. So that user experience is
more or less the same than writing or signing on paper, but you end up with
a high ,so secure digitally signed document. And the technical requirements
for using our software are very, very low, we offer them for all
platforms or major platforms, I have to say that means that
you can use them on Windows, on Mac iOS or on
Lennox and the, the new generation of software, the cloud suite gear software can also be used
on every device, which has a browser,
because they are we don't need any
local installation. We only need and normal an
actual browser so that you can have a connection to ours
with our cloud server. One of the important
things with our solution is in contrast to other
solutions is that you don't need to upload the documents to add to a cloud
server, and to get them signed. You only have these documents
on your own server in your own company domain, and the only information
which is leaving your company domain, which is sent to
the trust service providers, the hash value of the
document to be signed. And so there is no
lack of in information. There is no security hole in, in the installation where
you can have probably some problems in the future. If you upload your
secret and confidential documents to a
public cloud server. This is a short overview about
the remote signature with and the new flagship
product cloud suite gears? We have, they are on the
clouds with gear server. We offer the web interface and
they are also the interface for the certificates
and administration and for adopting the
APIs of the various trust service providers, We offer they're a
very powerful rest API. You can integrate very
easily into your own software environment. So if you don't want to take
the software as it is with our demo stack, you can
also integrate it very, very deeply in your own
software environment and your workflows, so that you end up with a seamless digital
document solution. And this is, I think
one of the best of us, we can do, and we have a lot of customers
who are especially looking for such a solution, which can be customized
to their own needs. And by preserving the fact that the documents will
never leave their own company domain (upbeat music)
