Cryptography - Blessing or Curse for the Information Society?

Dr. Seshu Bhagavathula

General remarks on cryptography

The technology that is used in the encryption of communication channels for sending messages through any electronic media is what has come to be known as cryptography. Consider a general case where some person wants to send a message through an electronic medium to another person and would not like the message to be heard or read by any other person. So what does he do to prevent someone from doing this? He employs, in general, some simple techniques like being cryptic about the topic so that the non-intended person – even if he reads - can not make much sense out of it. This is obviously effective only to a limited extent. But imagine that your message contains information that can not be cryptic any more as is the case with banking computers, but will contain all kinds of sensitive things that can are not allowed to be eavesdropped or tampered. This is where one requires some way of protection to the information being sent. In cryptography the message one sends is called the plaintext or cleartext. Encoding the contents of the message in such a way that it does not make any sense to outsiders is called encryption and the encrypted message is called ciphertext. The process of recovering the plaintext from the encoded message is called decryption. Encryption and decryption usually make use of a key where the decryption can be performed only by knowing the proper key, which is specific to the method used in encoding the message. Cryptography is the science, typically involving complex mathematics and cryptanalysis is the science that aims at breaking ciphers. Cryptanalysis, when performed by organizations, governments, institutions and individuals, is called science, depending upon the purpose, whereas it is called crime when the purpose is non-intended, i.e., aimed at taking undue advantage with damaging consequences to individuals, organizations or governments. Cryptography deals with all aspects of secure messaging, authentication, digital signatures, electronic money and other applications. Cryptography is the branch of mathematics that studies the mathematical foundations of cryptographic methods.

Until recently, cryptography was a subject of interest only to military applications, but as we move into information society, where millions of individuals (not to speak of other kind of users) making use of the electronic media for their everyday business, cryptography has become one of the most important tools for privacy, trust and access control in various forms of information transfer. Correspondingly, as more and more services are available on the electronic platforms all over the world, to be accessed from anywhere, abuse of the services has also dramatically increased.

The present scenario

At the heart of the whole issue are the hardware and the software that are used in protecting the information while transferring and in resident mode which in effect is to say that the computer security is essential. Emerging computer and communication technologies are rapidly changing the way we communicate and exchange information. Along with the obvious benefits like efficiency in communication, cost savings and so on, that come along with such a rapid development, there are inevitable challenges in keeping the information transfer safe and secure by using efficient cryptographic security techniques. Until recently there has been no real interest in the non-governmental sectors for encryption technologies. Electronic communications are now being widely used in the civilian sector since the "digital revolution" took place and cryptographic technologies have become an integral part of the global economy. Computers store and exchange an ever-increasing amount of highly personal information, including medical and financial data. In this electronic environment, the need for privacy enhancing technologies is apparent. Communication applications such as electronic mail and electronic fund transactions that can be in general termed as electronic commerce, require some form of encryption and authentication. As we all know governmental regulations of cryptographic security techniques endanger personal privacy. Encryption ensures the confidentiality of personal records, such as medical information, personal financial data and electronic mail. In a networked environment, such information is increasingly at risk of theft or misuse.

In some recent surveys aimed at finding out how much of cryptography is being used in various countries, the survey came up with surprising results: Most countries in the world today do not have any controls on the use of cryptography. In a vast majority of countries, cryptographic techniques can be freely used, made and distributed or sold without any restrictions, some which include even industrially developed countries. At the same time, one observes more and more relaxation of controls in the international law and policy, although it is not yet so far that development of market- based and user-driven cryptography products could be produced.

The strengths and weaknesses of cryptographic algorithms

While good cryptographic techniques should always aim at designing systems that are unbreakable, what is more important is – since unbreakable systems can claim their unbreakableness as long as they are not broken – the implementation of the systems properly with a complete and explicit documentation brought to the users on any possible mechanisms that can be used to circumvent the security. Theoretically, any cryptographic method with a key can be broken by brute force and with the availability of high-speed computers to try out all possible combinations of numbers, it is becoming more and more within the reach of organized criminals to have access to such systems. We have all heard the arithmetic that is involved in cracking the 32 bit keys which requires 232 steps a fact that this can be done by a simple commercial computer. Similar statements can be made of other keys also although computational power that is required to crack a 56 bit key will be quite enormous. But one must still keep in mind the kind of people or groups involved in such acts. It is obvious to imagine the backing such groups would have in terms of financial resources and the equipment they would have access to. However, the bit length of the key is not in itself a parameter that would decide the strength of the cryptographic system. Keeping the algorithm secret is one of way of making the system unbreakable, but in recent years we have seen people using reverse engineering techniques to break into systems, so this method can not be completely relied upon.

As the advances in cryptography brought us more complex and secure systems, cryptanalysis and attacks also increased correspondingly. Very often people describe cryptography products in terms of algorithms and key length, as if these are the only parameters that are synonymous to the amount of security a particular product offers or that these are the only factors that can be used to compare various algorithms like 128 bit keys mean good security or 40 bit keys mean weak security, etc.! But the fact is that it is not as simple because longer keys do not always mean better security. The non-intended cryptanalysts do not always try the brute force methods to barge into a system. They look for other weak points in the system especially in the implementation of the systems, because they know that many give too much of importance to the length of the key believing that it gives them the ultimate protection. Focussing on the cryptographic algorithms while ignoring other aspects of security is like defending your house not by building a fence around it, but by having only a very complex 10 pin door lock thinking that it takes 10 billion possible combinations for the thief to open the lock while the thief might simply crash in through the window! More often smart attackers devise other methods to circumvent the cryptographic security mechanisms.

There are various kinds of attacks known from the history of cryptography, most of which could have been avoided if proper attention is paid to its proper implementation of all the aspects of cryptographic systems. There are many attacks known until now while some concentrate on the cryptographic designs, the others concentrate on the implementations. Password hacking, attack against hardware, trust models, users, against failure recovery are some other known forms of attacks. While it is true that strong cryptographic systems rely on good cryptographic algorithms, digital signature algorithms, one way hash functions and message authentication codes, at the same time it is also true that if one of them is broken the system is also broken. These are powerful tools but making use of them in the right way is the key to guaranteeing security. There are cases known where the systems used powerful cryptographic algorithms but failed to check the size of values properly, or reuse the random parameters that should have never been reused. In some other cases, random number generators are found to be the source of weakness where cryptographic systems failed because good random number generators are mainly dependent on the particular hardware and software. How random is the randomness determines how secure the cryptographic system is. There were even cases where the generalization of random number generators has been shown to be the weak point, i.e., a random number generator good for a particular application does not necessarily mean it is good for all the other applications or systems.

As stated earlier, the implementation of the system is as important as having a strong encryption algorithm. In some cases it is reported that the plain text is not destroyed after it is encrypted, or others have used temporary files to protect the data loss during a system crash, or files written to virtual memory (hard disk space). It is also known that the infrequent memory refreshing caused retrieval of sensitive information later on and so on. Looking at these various cases reported, it is quite clear that there are other issues that also have to be considered apart from the cryptographic algorithms. Electronic commerce systems, for example, often make implementation trade-offs to enhance usability, like account reconciliation only once per day, or recording compromised keys on hotlists, etc. There are other areas which need to be mentioned here, like attack on passwords. Left to themselves, users don"t choose strong passwords - if forced, they don"t remember them. There are other limitations like only eight characters allowed as the password and converting them all to lower case and so on. Imagine how easy it is to find the right combination of eight-digit word than searching through a 64 bit random key. Hardware attacks have forced the companies to use tamper-resistant hardware for security like smart cards, electronic wallets, dongles, etc., but the tools against these also are getting better and better all the time. One of the examples of a successful attack against smart cards is the so called timing attack where some RSA private keys could be recovered by measuring the relative times the cryptographic operations took.

There is another form of cryptanalysis (attack) that comes from the pure signal analysis quarters. Connecting themselves to the cables that go in and out at suitable places around the building and analyzing the time and frequency (very often the signals are transmitted either in Time Division Multiplexing or Frequency Division Multiplexing) properties of the signals to derive the digits out of it is known to be another form being tried at. Measuring simply the voltages, radiation emissions, and other indirect electrical quantities would form the more basic levels of attacks.

The analysis on the weaknesses of the cryptographic implementations could go on forever. There are as many good cryptographic systems available as there are cryptanalysts. There are as many forms of attacks known as the number of cryptographic products known, each successively being upgraded and getting defeated. This is complementary to evolution where we don"t know who will survive at the end of the day!

So what is the future?

Where does this all lead? Sooner or later, every cryptographic system has the potential to be successfully attacked, probably in a completely unexpected way and with unexpected consequences. So it is important to anticipate and detect such an attack rather than bringing in new versions of software and hardware or designing longer keys. It is an old saying that says prevention is better than cure. Also important is to look at how does one contain the damage once the attack takes place. Are there ways that could bring back the system"s original configuration back? How can one detect while it is happening? Are there ways where one can produce evidence that could convince a judge and jury of guilt? The future of cryptography lies squarely on inventing systems that not only protect the users from known forms of attack "till to date", but also should provide security against all forms of attacks that the attackers might come up with in future.

Let"s look at some of the emerging technologies that might really put an end to the problem or at least that will give an indication of what is in store for us. Here we look at some of the concepts that could change the cryptography field by brining in some new inputs from an interdisciplinary thinking:

Immunological models for computer and communication security

Looking at the human body and its reaction to the virus attacks, one can derive some very useful information for cryptographic science. Human immune systems are remarkably effective in recognizing foreign antigens and the body"s own cells. The immune systems are highly parallel, distributive, adaptive and with learning capabilities. In addition, they are robust to attacks. The present cryptographic techniques are all meant to stop an intruder, but what can one do, when an intruder still manages to come in? One can look for answers in human immunology. The objective here is to exploit these attributes and derive mathematical models to show the feasibility of these techniques to security of computer systems and computer assisted transactions like electronic commerce, etc. One might ask where is the need for these Immunological Models? As computer and communication security is becoming increasingly important and complex, the time is ripe for looking "inwards" for a possible and a plausible solution. Human immunological models based security systems offer a better solution to the problem of in-security than the present fixed cryptographic techniques that are not adaptive and insensitive to change in the system. Apart from their distinguishing features mentioned above, being adaptive, they react to individual situations, that means security at a more fundamental level than the existing methods. For the readers of this presentation here, we give some of the names of models that are known from human immunology; cell activation and proliferation models, signal transduction models, immune networks, cellular automata based models, cross-regulation models, etc. There are many other models being investigated by researchers that offer alternative methods to the security of computer systems.

Self-healing networks – derived from the human body functioning

Present day cryptographic systems are fragile and prone to attacks that are irreparable. Human beings are involved in the care and repair of these systems at every stage in their operation. As the attacks become more sophisticated and the tools that will be used become more and more advanced technologically, it will be impossible for human beings to maintain them in future. Biological and social systems of comparable and greater complexity have self-healing processes which are crucial to their survival. It will be necessary to mimic such systems if our future cryptographic systems are to provide real security in complex and hostile environment. Here we touch upon some of the concepts that could be looked at for potential applications in the design of future cryptographic systems.

One could look at concepts like autonomous systems, where the maintenance can be given to a cryptographic immune system that exactly knows its state before the attack took place and has the necessary tools to put back all the code in place. This naturally requires elaborate methods in designing the audit trails and the process in which the data or program has been modified etc., things that are particular to different systems and not dealt with here. The implementation of such systems require all together new way of thinking which require all together different kind of tools to be developed, like preventive tool, infection or attack tool, biological protocols, computer lymphocytes, feedback systems and reactors, parallel information transfer, to name a few.

These models – naturally – have to be looked at from a system perspective and useful conversions have to be carried out so that they can be implemented in present day cryptographic systems. Imagine a cryptographic implementation that is equipped with a database with an intelligent program that looks at the pattern of the user activities (for that matter, any activity that happens in the system) to determine if the user is a normal user or a non-intended intruder. It not only identifies that it is an intruder but also can predict his next moves, thereby allowing the "cryptographic antigens" to do the rest. There is a lot of work going on in this field where researchers use various associative rule techniques coupled with artificial intelligence methods to predict the intentions of the attackers. This is perfectly possible because the number of moves that a non-intended intruder can make are limited because of the restrictions the cryptographic implementations put on the system. The buzzword could be "adaptive cryptography".

Quantum Cryptography

Quantum cryptography is a little bit away from the mainstream, but nonetheless worth mentioning for its uniqueness and potential to the subject on hand. There are universities and organizations that are looking at quantum mechanics to offer some alternative and better cryptographic techniques. With the emergence of semi- conducting and super-conducting devices as well as atomic and laser physics, we have seen quantum theory being applied to individual and physically larger systems. They say that it is now possible to envision the potential for quantum information technology, where machines would perform their operations and communicate according to the laws of quantum physics, as opposed to present technology, which is based on classical physics. Imagine a typical communication between two persons where quantum cryptography would allow them to exchange individual quantum systems to establish a shared random bit string used to encrypt the message. If they communicate using current encryption schemes, the correspondents can never be sure whether their message has been snooped upon, because classical information can be read without disturbing it at all. By using quantum cryptography, they can be sure of secure communication, because any eavesdropping will affect the quantum states. There are reports that indicate that already prototype installations of such systems are in operation that work over distances of about 10 km and with bit rates of about 10 kbit per second. However, it also mentioned that presently this technology is too costly to be brought into commercial use because of enormous investments involved.

Governmental policies needed

To allow the scientific community to really work and make products out of such research work, what is needed is a change in governmental policies. Restrictions on the development of new techniques hinder the freedom of thinking which will adversely affect the scientific community and consequently on the innovation potential of a science that is ready to give befitting answers to the present problem on cryptography. As a member of the research committee, the discussion on the policies are left to the experts that will decide the future of this field.