In February 2002 KPMG published a paper looking at the major issues and challenges as a result of the increasing use of cryptography to provide security and protection on the Internet. In particular it focuses on the need for better cryptographic key management policies, methodologies and technologies to avoid key chaos. Here, Daniel Murton, international marketing director at nCipher, who commissioned the paper, highlights the key points.
Cryptography dates back to Egyptian hieroglyphics circa 1900 B.C. and more recently played a vital part in World Wars I and II. Today, with the widespread acceptance of the Internet, cryptography is becoming widely adopted for mainstream applications such as e-commerce, secure e-mail, managing healthcare records, file and database storage, digital rights management and many more.
But all of the applications that rely on cryptography for security need to be underpinned by the proper and secure management of cryptographic keys. After all, you would not build a bank with expensive locks and safes and then leave the keys lying around for anyone to pick up.
A cryptographic key is merely data, a string of zeroes and ones that enable a cryptographic algorithm to scramble and unscramble information - a process known as encryption - or to electronically sign messages and documents as a means of proving their authenticity. At the root of cryptographic security is the secrecy of the keys used to perform these actions. Theft of a key potentially compromises the security of the whole system. The challenge for many organisations is to securely control the use of what might be hundreds or even thousands of these 'private' keys. The process of maintaining the integrity and authenticity of private keys is called key management and is widely recognised as a rapidly emerging administrative burden, one that many organisations are ill prepared for.
Keys have a complete lifecycle that needs to be managed from generation, distribution and storage to key use, back-up, recovery and eventually, termination or archiving.
Hardware vs software cryptography
Software-based cryptography is where the cryptographic algorithms, keys, cleartext data and ciphertext data all reside in the largely unprotected memory of a server or host computer. But because of the very random nature of cryptographic keys, they are easily spotted among other structured or normal unstructured data making them susceptible to identification, duplication, modification or substitution. The compromising of a private key can lead to risks such as eavesdropping, fake signatures, fraud and spoof IDs.
However, with hardware-based cryptography, the algorithm and keys are kept secure in the protected memory of a tamper-resistant device, commonly referred to as a Hardware Security Module (HSM). This ensures the confidentiality, integrity and authenticity of cryptographic keys and algorithms to reinforce the overall level of security.
As the demand for security and privacy grows, the use of encryption is becoming more and more widespread. The Internet is becoming increasingly encrypted. While only some 5% of Internet traffic is encrypted today, IDC supports the proposition that 100% of online business transactions and e-commerce activities will eventually be conducted over encrypted links. A more radical and optimistic view from Bank of America suggests that by 2005 all Internet traffic will be encrypted.
Whatever turns out to be the case, the sheer proliferation of cryptography will dramatically increase the number of cryptographic keys generated, distributed, installed, used and eventually terminated. Already today, major banks and corporations can have hundreds of servers and thousands of keys - yet it is likely that only a small percentage of these keys are managed effectively. That is if they even know where they are!
Organisations face the challenge of how to manage growing numbers of cryptographic keys used by multiple applications in multiple locations. A ten fold or 100 fold growth in the number of keys will dramatically stress the scalability of current key management procedures and will demand a new look at security policies that. Key management policies and practices must address the balance between operational requirements, device level security and environmental controls.
Thinking out of the box
While hardware-based cryptography is more secure than storing and managing keys in software, if keys are never allowed to leave the box, key management lacks flexibility and scalability.
One approach is to separate storage and operational functions within a combined hardware and software key management environment. Instead of always storing the application keys in the hardware security module, 'wrapper keys', which are used to encrypt the application keys are stored in hardware. The encrypted application keys are then stored as so called 'key blobs' on a host computer protected using triple-DES that is functionally unbreakable with any current or foreseeable technology.
While keys are never used or exposed unencrypted outside the HSM, storing and backing up outside of hardware securely removes keys from any individual point of compromise in the system. This ensures availability of keys whenever they are required across multiple servers and applications. And by storing keys in a totally secure format outside the hardware module, there is no restriction on the number of keys that can be managed and stored securely - providing a true scalable solution.
In addition to the growth in the volume of keys, the increasingly diverse nature of cryptographic applications means that the use and storage of keys is more widespread throughout the enterprise from mobile and desktop devices through network appliances and web servers, to back-end applications and databases.
Inevitably, one rule for key management does not fit all of these applications. Different applications have different threats and represent different costs if they are compromised. Instead, security policies and procedures need to be defined and implemented on a key-by-key basis depending on the type of application.
Furthermore, most companies will have cryptographic keys that are located across geographically dispersed sites and infrastructures. Currently, key management is a localised responsibility but with globalisation, the ability to securely administer cryptographic keys and devices remotely will become an important feature of any cross-border security architecture.
This will require powerful remote key management methods and techniques from either a single centralised site or multiple management sites to enforce policies and provide controls such as key separation and automatic key synchronisation.
Delegation of authority of automated systems
This same propagation and dispersion of cryptography implies that key management will also migrate from security officers with specialised skills and experience to disparate operational staff with more general knowledge and less appreciation for sound key management practices.
Therefore, coupled with remote key management capabilities will be a requirement for more automated key management tools. Such automation will lead to the use of software 'trusted agent' tools that will automatically execute processes and procedures in a secure, tamper-proof environment. In effect, the authority to perform a specific action such as authentication, validation or time-stamping is delegated to the trusted agent that may be located anywhere in the world.
The trusted agent may be developed by one company, installed at a second company and operated by yet another 'trusted' third party.
The ability to determine that adequate key management controls are in place requires regular reviews of policies, practices and procedures. And as more reliance is placed on automated key management and lesser-trained operators, the need for independent examination will increase. These will move away from traditional latent audits and migrate towards realtime auditing with online information feeds. The result is likely to translate into security scores being made available to the public. And maybe with more public awareness, companies will be pushed into adopting best practices.
Source: Daniel Murton, nCipher. NamITech, the secure solutions provider within the JSE Securities Exchange listed Nampak group has been appointed as lead distributor in Africa for nCipher.
23-25 April 2002
Grand Hall, Olympia, London, UK
Infosecurity Europe is Europe's largest IT Security Event. The show features a comprehensive range of FREE seminars and Keynote Sessions on the hottest information security topics as well as hosting the largest gathering of information security vendors and new products in Europe.