Freely subscribe to our NEWSLETTER

GS Mag: What are the different types of cryptography methods available on the market today?

Brian Tokuyoshi: There’s a couple of different ways to answer that question, but a good method to break it down are by the differences between symmetric and asymmetric cryptography. You could categorize the algorithms further by the type of cipher it uses, which are typically divided between block ciphers (an algorithm that encrypts a block of data and produces an encrypted output) and stream ciphers (which encrypts data flowing through the algorithm). For the purpose of this discussion, we’ll break the methods down by symmetric and asymmetric methods.

GS Mag: Can you describe the basic principles?

Brian Tokuyoshi: Symmetric cryptography is a fast method of cryptography which uses the same key for encryption and decryption. It’s used to transform information from plain text (i.e. information in its usable state) to cipher text (an encrypted state) and vice versa by the use an encryption key. One of the difficult challenges of symmetric cryptography is how to share the key with any other party that needs to access the data without exposing the key to a potential attacker. Any attacker who learns of the symmetric key would also be decrypt the data as well.
Asymmetric cryptography relies on a pair of keys that share a mathematical relationship with one another. One of the keys is known as a public key (which can be freely shared with other people) and the other is the private key (which is known only to the owner of the key and never shared), and each person participating in the process has their own pair of public/private keys. Asymmetric cryptography has the property that information encrypted to one key can only be decrypted by the other, so information encrypted to the public key can only be decrypted by the private, and information encrypted by the private key can only be decrypted by the public.

These principles can be combined together to perform a number of useful functions. For example, if the user wants to send an encrypted message to another party, the person can encrypt the data to the recipient’s public key, thus ensuring that only the recipient can decrypt it with the recipient’s private key. (Note that this is an over simplificiation of the process, because there is actually an additional step which incorporates symmetric key cryptography to improve the performance with most modern approaches to cryptography).
Another useful function is being able to demonstrate that a given user authored a document, and wishes to create a digital signature. The author could generate a hash of the document (which is somewhat akin to a checksum), and then encrypts the result using the user’s private key. When the recipient gets the document, the recipient decrypts the author’s hash result, performs the hash of the document and compares the results. This process can check to make sure that the information has not been tampered (in which case the hashs would not match), and it can also prove that the document is not forged (since only the possessor of the private key would be able to create a piece of data decryptable by the author’s public key).

GS Mag: How are the primary uses of these different types of methods?

Brian Tokuyoshi: Symmetric key cryptography is used in systems which require fast encryption/decryption of data.

Asymmetric key cryptography is largely used within a system known as PKI (which addresses the problem of proving who owns a particular public key), and which enables a number of security features such as strong authentication (with smart cards, x.509 certifciates in networks), network encryption (VPN, SSL and TLS), and digital signatures.

Note that symmetric key cryptography and asymmetric key cryptography are often paired together to perform encryption, such as with PGP or network protocols such as TLS. These systems use a symmetric key in order to perform the actual encryption, due to its speed. As mentioned before, the challenge behind symmetric key cryptography is finding a safe way to give the key to the recipient of the data. With PGP, the sender encrypts the symmetric key with the recipient’s public key, thus assuring only the recipient can decrypt the symmetric key and recover the information. This method leverages the advantages of both environments, i.e. using symmetric key for fast encryption/decryption and the asymmetric keys to exchange the symmetric key. It also has the advantage of keeping file sizes small when being sent to multiple recipients, because each additional recipient only increases the size of the file by the size of another copy of the encrypted symmetric key. If the file was encrypted solely with asymmetric key cryptography, it would not only take considerable time to encrypt/decrypt, it also would double the size of the file for each additional recipient (since each recipient is only able to decrypt their own copy of the message).

GS Mag: What are the main drivers for the development of cryptography in France?

Brian Tokuyoshi: Like many countries around the world, the driver starts with the problems of controlling access to sensitive information. Plaintext information, once stolen, is subject to a number of potential problems, such as being used to commit fraud and violation of privacy laws. In the past, it used to be good enough to protect the perimeter of the network to prevent bad guys from getting inside. Such practices are no longer good enough and protection needs to go around the data itself, which has been the driver for the additional of data encryption as well as all of the additional security services provided by cryptography. It’s about establishing more data privacy to prevent data breaches and meet compliance requirements, along with establishing more trust in the data to know that it’s reliable and has not been tampered.
Cloud adds an additional twist to the mix because now the perimeter is irrelevant. Data protection is the only measure to keep data private from all 3rd parties, including the cloud provider’s system administrator.
Perhaps the best answer is that it should be used every where data needs to be protected, but only up to the ability of the company to manage it. Poorly managed encryption poses just as much a threat to the loss of data, especially within the realm of managing the encryption keys. Unfortunately, many companies do not have a choice of how much encryption to use – they are being pushed by compliance requirements to encrypt more information, and often they have a lack of tools to manage the keys, and this combination is creating tremendous pressure on the enterprise as they get squeezed in the middle.

The better approach is not to think about encryption in terms of the application as the primary consideration. What’s needed is to think about the requirements for management, and layer as much of the encryption on top of that management framework in order to prevent future blunders.

GS Mag: Can one or should one encrypt everything in a company?

Brian Tokuyoshi: One should think about encryption as a tool, and it’s important to consider how to use a tool properly to achieve proper goals. Just like an axe is a tool, one that can both help you build things as well as break things down.

Encryption should be used in proper measure, and it would be oversimplied to say that all data in a company should be encrypted. The reason being is that poorly managed encryption and encryption that interferes with workflows and employee processes can be counterproductive.

With that said though, most companies do not encrypt enough data, but they are also unprepared to encrypt more. That’s why the concept of managing encryption is so important – so that organizations can expand their use of encryption without creating an administrative burden or introducing a self-created data loss incident.
 How do you deploy cryptography solutions?

The ideal scenario is to deploy management services before deploying any cryptography. The problem that many organizations encounter down the road is that each cryptography solution has its own management tools, and having multiple management tools creates an administrative burden that can be difficult to recover from.

Think about planning administration, key management, logging, and policy before thinking about how to actually encrypt the data itself. The cryptography portion of the equation is easy – there’s software to do it. The management is far more difficult – it requires staff to do that.

GS Mag: Where do we stand with email encryption?

Brian Tokuyoshi: Email encryption has gone through several waves on its way to a mainstream technology. It was probably one of the first mainstream uses of cryptography, and to this day there are still a couple of different concepts that one should consider before making a decision:

1) What are you trying to protect? – some solutions only protect certain mail streams and not others, and even the adoption of encryption won’t necessarily address concerns about separation of duties.

2) Who do you need to talk to? – there are some email encryption solutions that force the recipient to use your mail environment. Being able to have choices and to interoperate with the encrypted email solution that a partner has without requiring them to rewire their environment is an important consideration.

3) Where are you accessing email? – email needs to be accessible from a number of locations, including on the road and on mobile devices. If they are accessing secure email from a protected device, it should be accessible in the native interface of the device, and not require jumping out to a separate web browser.

GS Mag: What are your recommendations in this area?

Brian Tokuyoshi: Just like any prudent approach to a long term project, a proven solutions with support for open standards is the key to a sustainable solution. There are too many ways to get email encryption done wrong, especially with shortcomings seem by being forced into one type of architecture or being limited to proprietary encryption methods. Open standards is the only guarantee that you have full control over your own data.

Another consideration is to look at the enterprise methods for data recovery and what tools are in place to support ediscovery and investigations. There are too many ways to hit a dead end if these considerations aren’t thought out beforehand, before any data gets encrypted in the first place.

GS Mag: What are the main issues with current cryptography methods?

Brian Tokuyoshi: As mentioned before, key management is a pressing issue that many organizations face as their environment grows. This will continue to grow to become a bigger problem in the future as more cryptography gets deployed throughout the enterprise.

The second problem is the usability of many of the cryptography services – end users don’t want to use things that interfere with their work, and many cryptography solutions that proclaim to be easy to use actually require extensive changes to their workflow. Users will try to find shortcuts around things that they see as impediments to their work, because too often, cryptography software is often written to prevent something rather than enable something. Users feel like they are losing freedoms rather than gaining freedoms from knowing their data is secure. Improving the usability and reducing the impact on users is important in this area.

GS Mag: How do cyber-criminals use them?

Brian Tokuyoshi: It’s actually quite rare that criminals attack the cryptography itself when it comes to standards based cryptography. It’s still far too difficult to brute force a piece of data unless the actual data itself is worth the investment in time and resources.

With that said, criminals are attacking the human element of the equation – instead of going after the secure network connection, it’s easier to insert malware on the user’s computer to steal the raw data before it goes on the pipe. The same goes for breaking into encrypted drives – it’s easier to brute force the user’s passphrase that protects the symmetric key than it is to break the key itself.

With that said, encryption by itself is not a complete security solution – but it’s one that‘s been missing in the past. Before, people used to rely just on their firewall for security and didn’t use encryption, and that’s been exploited. In the same way, firewalls never went away, they still have a useful purpose, but encryption by itself doesn’t replace the need for a firewall. Both need to be put together in order to prevent the different avenues of attacks.

GS Mag: How can we resolve these issues?

Brian Tokuyoshi: Encryption by itself only talks about protecting the data. One probably also needs to think about the methods for using cryptography to improve other security services, such as using signatures to check the fidelity of the data (to prevent forged data from causing fraud scenarios) or using strong authentication to get around the problems of the passphrase brute force exploits. These are just a small number of the issues at play, because we expect to see Man in the Middle attacks continue to grow, but yet again, this is an area that can be solved with strong authentication (using x.509 certificates. Yet businesses tend to avoid deploying these solutions in the past due to the difficulty in managing the keys. With modern key management solutions, we expect to see the adoption of such technologies go mainstream.

GS Mag: In your opinion, will use of cryptography become more widespread?

Brian Tokuyoshi: Yes, unquestionably. We believe that in the past, there was a general assumption that the unencrypted state of the data was the normal state, and the encrypted state of the data was only temporary until it is decrypted. In the future, these roles will get switched around, because the encrypted state of the data is the normal state, and it’s only encrypted when authorized people need to use it.

GS Mag: What will the cryptography of the future look like?

Brian Tokuyoshi: The big news in cryptography tends to be around the upper strata – how many billions of years would it theoretically take to break this key, or what kind of threats exist in certain algorithms in certain conditions. To businesses though, these issues tend to be largely theoretical and what really drives their interest is how to make cryptography more practical. We’ve seen how compliance has made encryption mainstream with a lot of businesses, but we’ve also seen that many businesses only encrypt enough to protect them from an audit, not to necessarily protect themselves from potential data loss threats coming from different directions. So the future of cryptography will come from more uses of mainstream cryptography being adopted into a system that makes it part of the very fabric of all the products they normally use. For instance, people don’t even know they are using cryptography with their smart card badge, yet it provides a number of security services and it solves some major problems dealing with identity theft and corporate security. So the old photo employee badge has transformed from a relatively dumb device (piece of plastic with a photo) to something that embraces cryptography. We expected the transformation to apply all across the IT environment, whether it’s computer systems or network, cryptography will be something that is baked in rather than bolted on.

GS Mag: What will the arrival of SHA-3 change?

Brian Tokuyoshi: In short, it will change everything and nothing. It will affect everything because the adoption of SHA-3 will lead towards rapid uptake in a broad range of cryptographic operations. We know that there are current attacks on the existing hash algorithms that led towards the working group for SHA-3 and the need to get a new standard in place before the current one falls.

Yet it seems that even with known attacks on existing algorithms, there has been rather slow adoption towards fixing the problems until a significant security incident forces an emergency response, when a more measured response would have been more prudent. As a result, people will probably do nothing for SHA-3 for a long time. Let’s take MD5 for an example – the attacks on MD5 started showing up in the mid 90s, and started to become more serious in the early 2004 time frame. However, it wasn’t until the attack on MD5 in 2007 before it forced everyone to start moving away.

Likewise, NIST is updating its guidelines on key strengths and algorithms in order to phase out older, weaker keylengths. It would be unthinkable that anyone is still using those older key lengths, but yet it took NIST to push a standard in order to get the process to phase it out.

With that said, the cryptography is still significantly strong and it won’t fall easily to attackers of average ability. Yet even though when it would be easier to do a phased adoption of better standards, there’s a slow adoption curve due to how many of the keys exist and how difficult it might be to switch. It took an emergency response to push out use of the ailing MD5 algorithm. At that time just prior to the attack, people weighed the cost/benefit ratio and thought that leaving the hash function in place was less expensive than replacing it, even though the potential for a new attack loomed on the horizon.

What’s needed though is better tools to manage keys/certificates and rotate them out so that in the event of a rotation and new standard, there’s better tools to help make the transition. That’s what made the past rotations so difficult and slow, and why people do nothing even when something is better. With better key management, hopefully these past experiences will be better handled in the future.

We’ve seen the impact of the attack on MD5 and how it led to 3rd party certificate authorities to reissue their certs on a new root in order to prevent the potential