Method and device for guaranteeing the integrity and authenticity of a set of data

<img img-format="tif" img-content="drawing" file="IL147589AD00021.tif" id="idf0001" />

METHOD AND DEVICE FOR GUARANTEEING THE INTEGRITY AND AUTHENTICITY OF A SET OF DATA omm nonvD bw mmnNi ivw nnunnb prim nvnw 1 147,589/2

A METHOD AND DEVICE TO GUARANTEE THE INTEGRITY AND AUTHENTICITY OF A SET OF DATA

This invention concerns the area of the control of the integrity and authenticity of data, and in particular the downloading of software.

The invention is applied to all those apparatuses that contain at least one central unit such as those currently used in information technology, that is to say, with a processor that has at least a part of its program inside a multiple rewrite memory.

It is well known that the alteration or the damage of data leaves traces in certain parts of the information processed and stored in a memory, either before or after being processed. It is also known that a simple mathematical technique such as “checksum” is used in order to determine if the data taken into consideration has been modified by establishing a checksum reference.

However, most checksum systems have drawbacks. Thus, during the course of mathematical operations, the propagation of compensatory aleatory errors may occur, compensating each other and giving an identical result to the one expected. Consequently, verification by the known methods will be inoperative in certain cases.

There is, therefore, a problem that has not been solved in a satisfactory manner, which consists in improving the reliability and the security achieved by said checksum operations, particularly in data sent along public routes and destined to be stocked in a memory program.

The scope of the present invention is reached by adding to the data to be sent a control block comprising the result of an operation done on all or a part of said data, a mono-directional and clash-free operation.

Within the sphere of this invention, a message is a set of data destined to be updated or the loading of a software. 2 147,589/2 “Mono-directional” means an operation that is easy to perform in one direction but almost impossible to perform in the other direction. For example, the operation XY is easy to carry out, while the operation Y&gt;/X is much more difficult.

The term clash-free operation means an operation according to which no other combination of data that is entered gives a similar result.

Within the sphere of this invention, this mono-directional operation is a mathematical application H of a source group towards an object group, in which each x element of the source group is attributed with an H(x) symbol. These functions are particularly useful when there are functions known as Hash, as they are defined on page 27 of the RSA Laboratories publication “Frequently Asked Questions About Today’s Cryptography, v.4.0”. The x element can be of any length, but H(X) always has a series of characters of a fixed length (fixed-size string). Such a function is difficult to invert, that is to say, knowing H(X) does not generally mean that x can be found. What is more, it is said to be clash-free when it is injective, that is, that H(Y) = H(X) leads to y = x, like Η(Υ) Φ H(X) leads to y Ψ x.

The present invention consists in replacing the known checksum operations by using mono-directional functions, in particular Hash functions.

In a particular embodiment, the group considered includes a transmitter part, located inside a control centre, and a receiver that can be made up of quite a large number of peripheral units, which work in a similar way. The aim is to guarantee that the software sent by the transmission part is received in an authentic and complete way by each of the peripheral units. In line with the terminology used in pay television, which represents an important but not exclusive application of the invention, the peripheral units will be called IRD (Integrated Receiver Decoder) in the following part of the paper, and include a receiver, a decoder to process the signal received by the decoder, a central processing unit, or CPU, that works preferably with a non-volatile memory, as well as various peripherals. A non-volatile memory is a memory where the contents are maintained intact even if the main current supply is cut, for example through at least one 3 147,589/2 independent source of energy such as batteries. Other types of non-volatile memories can be used, such as EEPROM, Flash EPROM or FEPROM. It is this non-volatile memory that keeps the data safeguarded in case of an interruption in the supply of current. It is essential for making the IRD processor work well.

The information is received by the IRD coming from the control centre, in the form of a stream of data arriving at the receiver of the IRD unit. In the case of coded television, or more in general interactive, the stream of data includes video information, audio information, data information, executable applications like “applets” and, finally, various types of data check information.

In this case it is a question of guaranteeing that the information is received correctly and interpreted by the IRD before being stored in the operational memory, particularly the execute data, that is, the software.

The receiver of the IRD transmits them to a decoder, which then puts them in circulation in the IRD by means of a bus. Connected to the bus there are a specialised multimedia processor that is, in turn, connected to a monitor and to one or more loudspeakers, the aforementioned non-volatile memory and one or more optional sub-devices. It is the IRD processor (CPU) that manages and controls the correct operation of the latter, as well as the different sub-devices, such as a test channel, an interface for chip card, an auxiliary memory pack, other processors or a modem. What is more, the control centre may receive exchange information, for example through a modem connected to the public telecommunications network.

These sub-devices themselves could be the source of errors that have to be detected and corrected, especially in the case of the loading of a new version of IRD operating software and particularly of its CPU, or of certain execute programs for the IRD or its components.

The software and the data for which the authenticity and integrity must be guaranteed may be loaded by various means. One of these means, as has already been said, consists in sending to the aforementioned receiver, identifying it in a way recognizable for the central unit, an update of the memory with the stream of 4 147,589/2 data, including a number of data blocks Ml, M2, ...Mn to allow the data Ml to Mn as well as a heading which allows to identify the data Ml to Mn.

Alternatively, or as a supplement, the data blocks may reach the IRD through one of its optional sub-devices such as the modem, for example.

The data blocks Ml, M2, .. .Mn may be sent in clear without any drawbacks, that is to say, without yet being ciphered at this stage, within the sphere of this invention.

The method according to the invention consists, in this form, in applying first of all, during the transmission stage, a mono-directional or Hash function to a part or to all of the data blocks Ml, M2, ...Mn to obtain a representative Hx result of the Ml to Mn group. The data blocks Ml to Mn may be processed separately just the same and obtain the Hxl result corresponding to Ml, Hx2 corresponding to M2 and Hxn corresponding to Mn. This or these Hx result(s) are put into a control block, which may be a part of the heading at a non-aleatory place. When several Hx results are produced, which are representative for each Mn part, they are placed in different sites in the control block. The control block also comprises IS service information like the version of the Mn data, the date of start-up or the validity duration of this data. A particularly crucial sector for the authentication of the data concerns the systems by which the data is transmitted along public routes such as Hertzian-wave, telephonic or internet routes. In this case, an intruder may take the place of the control centre and send data to modify the operation of the target system.

This invention also extends to a method for guaranteeing the authenticity of the transmitter of the message by transmitting the whole or part of the control block in the form of a cryptogram. In one embodiment, this cryptogram is calculated on the whole of the data of the control block. The message is composed by an identification part in clear to announce the type of message arriving at the peripheral unit, by a signature representing the encrypted control block and data Ml to Mn. In the same way as for the calculation of the function Hx, this signature can itself decompose into several signatures (Sxl, Sx2 ... Sxn) on 5 147,589/2 different parts of the control block. For example, the information relating to the data Ml, i.e. the result Hxl as well as the associated services data IS1, are encrypted to form the cryptogram Sxl. This algorithm of encrypting can, in a known way, comprise the use of an asymmetrical key in private form, that is to say using the private key of the RSA algorithm.

The only thing that remains to be done is to assemble the identifier block, the control block in its clear form (without function of encrypting) or in its encrypted form, to the proper data Ml, M2...Mn, and to send this message towards the IRD(s).

The reception part of the IRD carries out the temporary stocking of the message and, if the control block is encrypted, the IRD determines the control block in clear by applying to the encrypted block the algorithm of decoding using the public key of the RSA algorithm.

Once the control block is available in clear, the IRD applies the same mono-directional function as in the moment of the transmission on the blocks of data Ml to Mn to determine the result(s) Hy of the Hash function, results which will be compared to the data contained in the control block.

If the comparison Hx=Hy is positive, one can affirm that the data Ml to Mn have not been corrupted or modified during the transport, and, if the operation of encrypting the control block has been applied, one can furthermore affirm that the data are in fact by the author from who this data is expected.

In this case, the operation is considered as successful and the new data can, either immediately at a date agreed on, be transferred from the temporary memory (Ml to Mn) to the operational memory (Ml to Mn).

In a particular embodiment of the invention, the control block also includes one or more certificates delivered by an external authority to the control centre. By analogy to the results of the hash functions on the data, the certificate Cx can itself decompose into a certificate Cxi corresponding to Ml, Cx2 corresponding 6 147,589/2 to M2 ... and Cxn corresponding to Mn. Cx means the whole of these certificates in the following description. Examples of methods of determining an authenticity certificate are described in the documents U.S. 5,373,561; U.S. 5,136,646 and U.S. 5,136,647.

In a particular embodiment of the invention, these certificates comprise a piece of information of a digital time-clock or “Timestamp”. This certificate has the advantage of doubtlessly representing the whole of data Ml to Mn because these data have been presented to an external authority, which in return delivered information called authentication information allowing to trace back the origin of the data. This presents the advantage to allow to situate the origin of an error in an unquestionable way, because proved by a recognized authority, as well as, from the legal side, to prove the non-implication of this or that part of the procedure in the detected error. This is important when the error can have juridical consequences.

The force of the method lies in part in the quality of the mono-directional function H and in the encrypted signature of the latter. So, a simple checksum does not allow to detect the exchange of two blocks of characters in the data because the addition is, in mathematics, known to be commutative and associative. On the other hand, a “hash” function result Hx is a very realistic image of x, even if x is much longer than Hx. If exchanges of characters are carried out in the string of characters x, the function H(x) will detect it immediately, and the system can not work following this detection. This results in an increased security.

An important aspect of the invention is that it allows to verify in any moment the validity of the data memorized in the peripheral unit. In fact, the presence in memory of this information of control allows the unit to proceed to . a reliable self-check. This verification gives a result without comparison with the checksum usually applied on the program memory. If this control gives a result different from that of reference, the unit disposes of means (connection modem, 7 147,589/2 connection on cable channel) to inform an external entity, for example the control centre, about the non-compliance of the programme.

If the preferential means of the invention for the generation and transmission of the control information is a control centre, the invention covers a peripheral unit whose program or part of it is initially loaded with the control information as described above. This can be carried out in a production site or in the moment of the initialisation preceding the sale by means of the processor, or by distance downloading of this control information by one of the peripherals in an initialisation phase.

This invention will be understood more in detail thanks to the following drawings, which are not limiting, and in which:

Figure 1 represents the mechanism for the preparation of the data sending the message; and

Figure 2 represents a block diagram of an IRD.

In Figure 1, the data composing a software is divided in blocks of data Ml to Mn. This data are treated by the hash function so as to determine the results Hxl to Hxn, which are representative of the whole of data. On this diagram, the optional part, which consists in adding, certificates Cxi to Cxn to the results Hx is represented by the block AE, which means external authority. This external authority returns the certificates Cx. The control block CB is then composed of the results Hx and the certificates Cx. Service information IS characterizing the data are also joined.

In the embodiment including the cryptogram Sx, the control block (CB) previously composed is treated for encrypting, by means of the private key PrK, the control centre. 8 147,589/2

The final result, as disclosed in Figure 1, represents the message ready to be sent and started by the message identifier Id, the encrypted control block Sx and the data Ml to Mn.

In Figure 2, an IRD or Integrated Receiver Decoder is represented, constituting the peripheral part of the plant to which is applied the method according to the invention in the way of realization described hereafter. This IRD includes a central bus DB to which all the different modules connect. The central module of the IRD is constituted by the processor CPU, which has the task to carry out the different treatments. A receiver REC receives a flux of data comprising video information, audio, data and executable applications through supports as varied as cable, a hertzian antenna, a satellite dish, Internet or other known technologies. This receiver REC is linked to an interface DC, itself connected to the bus DB.

To this bus DB are also connected: - a multimedia MP processor specialized in the treatment of video or audio information, that it directs respectively towards a visualization screen VD and loud-speakers AD; - a test channel TC, itself being able to be linked to a tester TEST serving for factory settings and maintenance; - a non-volatile memory NVM, made independent from the main feed by its own feed source; - an interface INT for microchip card, physically receiving a microchip card SM; - an auxiliary or still bulk memory TMEM; - a modem MD, connected to the public network NET, which adopts known technologies and supports; 9 147,589/2 - other processors OP, DP assuming diverse functions according to the user’s wish, especially that of data processing (Data).

The CPU controls the software updatings, an example for which is described overleaf. It accepts them or rejects them according to the result of the tests performed according to the method, which is the object of the invention.

These versions of software of the CPU of the IRD can reach the IRD by the receiver REC, by the tester TEST, by the microchip card SM, or by the network NET. Later, the case where they arrive at the IRD through the receiver REC will be described, with the video and audio information flow. A set of data, representing a new software version arriving at the IRD, is stocked in the memory of the IRD, with the service information, after having been controlled for its authenticity and its integrity. This allows the control centre to load this software version in a great number of peripheral IRDs, and to begin its start-up without error by the whole of these IRDs.

The invention does not limit itself to the examples mentioned above and extends to other forms of realization. For example, in the control block it is possible to transmit the results Hx and the cryptograms Sx. Therefore, the operations of integrity verification can be independent from the authenticity verification operations. The order of the information as described in Figure 1 is indicated as a not limiting example and can be any. 10 147,589/2

It should be noted that the verification information of the control block CB such as the code Hx and the signature Sx, are memorized in the IRD for future use. In this way, a validation or confirmation of the integrity and the authenticity of the software can be asked from the IRD at any moment, by the microchip card SM, the tester TEST, or on request arriving by the network NET or the receiver REC. In this case, the IRD will recalculate these different parameters Hy and will compare the latter with the memorized reference values. Similarly, the CPU too can, at regular intervals, start this verification procedure.

In another embodiment, the verification is carried out not by the CPU, but by an external entity such as the tester TEST, or by the network NET or the connection REC. In this case, all or part of the data Mn are transmitted for verifications of the control information Hy, which is compared with the original Hx values. A supplementary verification can be made with the value Hx memorized by the IRD.

According to an embodiment of the invention, the verification information (Hx, Sx) is memorized in one of the peripheral units, for example the microchip card SM which is thought to be inviolable. Therefore, it is not possible to modify Hx for making it correspond to Hy, despite the big differences of the data Ml to Mn.

For the encrypting function, it is recommended to use an algorithm by using an asymmetrical key in the private form PrK. This is for example the case of the algorithms of the RSA type (Rivest, Shamir &amp;amp; Adleman). This algorithm is applied on the control block only for practical reasons. One could apply it also to the whole of data according to the time available for this operation or the length of the data. 11 147,589/2

Once the message has been received by the IRD, the message is decomposed and the different elements are stocked in the temporary memory TMEM. The IRD applies to the blocks of data Ml to Mn the same treatment as in the moment of the transmission but in reverse order. It is clear than when one receives the control block in encrypted form, the first operation consists in decoding the control block by the public key PuK to obtain the control block in clear CB’.

The following step consists in carrying out the mono-directional function H on the data Ml to Mn with the values Hyl to Hyn as results. When an error enters in the memory blocks Ml, M2, ...Mn during the transmission of the message, this error itself influences Hy which is then be different from Hx contained in the control block, and the data Ml to Mn will be rejected. So, obtaining the values of reference Hx by an encrypted function, added to the calculation of function H on the data, allows to guarantee the integrity and the authenticity of the data Ml to Mn.

As examples of functions H, the functions MD2, MD5 and SHA-1 are known.

Twa , crrixan rwzn ara inia^a pnow pnizn irn nr qaoa ,ρνιη unan lasana mavia na^maa np’ioa .zruwan rwaa mp^an pmi1? axnna □ιηπη Pi? _____88°4003 01 J' ·aoi 085122+0200 .(mcna navin) cras&amp;amp;’an rwa