Sunday, July 19th, 2015

What Is A Signature — Really?

Signing an agreement often means to write your name with ink on paper. At least in the Western world. A quick search on the Internet reveals that stamps have been used for similar purposes in many cultures since ancient times. Personal stamps are still in common use, for instance, in Korea and Japan. Thousands of years ago, an agreement could be made in two copies. One was permanently enclosed in a clay jar. The only way to access the sealed copy was to break the jar.

What is the purpose of these various customs? How do they compare to modern-day digital signatures?

Let’s start in the digital end. The big difference between digital and physical signatures is that the distinction between original and copy has disappeared. Keep copying digital documents, and all you produce is a number of originals. This is important because many of the techniques used for physical signatures are meant to certify that a piece of information is the original.

The modern-day buzz words for the defining properties of a signature are,

  • Authentication: Prove that the document originates from the signer and no one else
  • Integrity: Prove that the document has not been modified after signing
  • Non-Repudiation: Make it difficult for the signer to deny that they signed the document

Prove to whom? There is an underlying assumption that a third, independent party may be called in to resolve disputes over an agreement. It doesn’t have to be a court, but courts set the highest standards when it comes to assessing evidence, so let’s assume that the signature should stand up to court examination.

Authentication in the ink and paper world is based on the assumption that your handwritten signature or personal stamp is unique. Really important documents, like wills and real estate transfers may require witnesses to confirm that the signature was made by the signer in person.

Digital authentication (given current technology) relies on a personal certificate embedded on a smart card. It means that the signer was in physical possession of the card and knew the PIN code required to unlock it. You check authenticity by decrypting the signature using the public key of the signer.

Integrity of paper documents is commonly achieved by letting all parties have a copy of the document. In certain cases, like real estate transfers, an extra copy is registered with an official archive. Any changes to the original document would be discovered by comparing it to the official copy. This is, in fact, also the purpose of sealing a copy in a jar.

The integrity of a digital document is guaranteed by taking a checksum of its contents. The checksum is made part of the signature. It is incredibly difficult to make a change to the document without altering the checksum.

Non-repudiation of paper documents follows the same logic as authentication. An ink signature may be forged, but it takes a determined effort. Add witnesses, and the signer will have a hard time claiming they didn’t sign.

In the digital world, non-repudiation relies on public key encryption. If a signature can be decrypted using the public key of the signer, then you may be certain that the signature was made with their smart card.

Validation of a digital signature means to examine it in its context, that is, with the document it belongs to. The three key properties, authentication, integrity and non-repudiation, depend on the signature being valid.

Legally binding signatures, at least in Europe, must be based on certificate technology that, in turn, is based on public key encryption technology. An important purpose of certificate technology is to establish a chain of trust that ends with a Certificate Authority (CA), the trust anchor. Certificates use public key encryption in order to work around the difficult problem of communicating secret keys. A CA shares its public key as publicly as possible. For instance, every web browser around the globe includes a few hundred CA public keys.

Validation of a digital signature begins with following its chain of trust to the trust anchor. To validate the trust anchor you must be in possession of its public key.

The next step is to verify that the signature contains the checksum of the document.

The beauty of this procedure is that it can be performed by an independent third party. It is also subject to international standards and built into many modern programming languages, like Java.

In real life, there are two catches.

  1. Most digital signature implementations deny the signer (or anyone else) access to signatures. You can’t validate a signature you don’t have.
  2. In Sweden: The main Swedish provider of smart cards does not publish its (so-called) public key unless you pay for a commercial contract. Without the CA public key the chain of trust ends in empty space.

Comments are closed.