Classifications

• • G—PHYSICS
  • G06—COMPUTING; CALCULATING OR COUNTING
  • G06F—ELECTRIC DIGITAL DATA PROCESSING
  • G06F9/00—Arrangements for program control, e.g. control units
  • G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
  • G06F9/46—Multiprogramming arrangements
  • G06F9/465—Distributed object oriented systems
• • G—PHYSICS
  • G06—COMPUTING; CALCULATING OR COUNTING
  • G06F—ELECTRIC DIGITAL DATA PROCESSING
  • G06F9/00—Arrangements for program control, e.g. control units
  • G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
  • G06F9/46—Multiprogramming arrangements

Definitions

• the present invention relates to a method and system for processing a request of a customer. More particularly, the invention relates to cryptographic security of mobile agents.
• mobile code is a programming paradigm that becomes more and more important. It provides a flexible way to structure cooperative computation in distributed systems.
• the Internet is full of mobile code fragments, such as Java applets, which represent only a simple form of mobile code.
• Mobile agents are mobile code that acts autonomously on behalf of a user for continuous collecting, filtering, and processing of information. They combine the benefits of the agent paradigm, such as reacting to a changing environment and autonomous operation, with the features of remote code execution; they operate in computer networks and are capable of moving from server to server as necessary to full their goals. Important applications include mobile computing, where bandwidth is limited or users are disconnected, data retrieval from large repositories, and configuration management of software and networks. The vision of mobile agents roaming the Internet may soon become reality as the paradigm is incorporated in large-scale applications.
• Mobile code is to be understood as a program that is produced by one entity, called the originator, and is subsequently transferred to a entity, the host, immediately before it is executed by the host. In other words, no manual intervention, such as performing an installation or running a setup routine, is required on behalf of the host; mobile code comes ready to run.
• mobile agents are capable of continued, autonomous operation disconnected from the originator and migrate freely to other hosts during their lifetime. Such agents have also been called itinerant agents.
• Mobile code is exposed to various security threats: a malicious host may examine the code, try to learn the secrets carried by an agent, and exploit this knowledge in its interaction with the agent to gain an unfair advantage. A host might also try to manipulate the result of a computation.
• a further form of code is active mobile code that performs some immediate action on the host. Thereby often information about the encrypted computation is leaked to the host whereby only the originator shall receive any output.
• US Patent No. 6,026,374 is related to a system and method using of a trusted third party to provide a description of an information product to potential buyers without disclosing the entire contents of the information products, which might compromise the interests of the seller. The buyer trusts the third party to give an accurate description of the information that is for sale, while the seller trusts the third party not to reveal an excessive amount of the information product's content.
• the system can include a seller of information products, a buyer of such products, and a trusted third party summarizer, each operating as a node in a communications network, such as the Internet.
• a disadvantage of this system and method is that the third party has to be a trusted one and that this third party gets information and learns about everything. This could be dangerous if said third party gets cracked. Moreover, several messages are necessary to process the request of the buyer.
• the invention provides a method and system for processing securely an originator request of a customer, i.e. the request being initiated by the customer.
• This originator request is sent within a mobile code or agent to at least one first entity.
• the method for processing the originator request comprises the steps of a) sending from the customer, i.e.
• the originator request to the or each first entity; b) connecting the or each first entity to a computation entity; c) adding by the or each first entity, on receipt of the originator request, information concerning the originator request thereby forming a first-modified request; d) sending at least part of the first-modified request to at least the computation entity; e) having received at least part of the first-modified request by the computation entity deriving a computation-entity result from the at least part of the first-modified request; f) sending at least part of the computation-entity result to the or each first entity; g) having received at least part of the computation-entity result by the or each first entity deriving therefrom a first-entity result and forwarding it at least in part; and h) having received at least part of the first-entity result by the customer deriving therefrom a customer result.
• the mobile code or agent comprising the originator request is sent around a network to several entities whereby the code or at least fragments thereof can be securely executed without any additional client hardware at the first entity.
• Security is obtained through the computation entity that is a generic independent entity.
• This independent entity can be a computation service that performs an encrypted computation on behalf of the mobile agent, but does not learn anything about the encrypted computation.
• the independent entity can serve many different applications whereby nothing about its usage is necessary to know before deploying it. Privacy as well as authenticity for the mobile agents can be achieved.
• the computation service itself does not learn anything about the computation, assumed it does not collude with the code originator or the first entity.
• the independent entity may be universal rather than bound to a particular service or to an application context.
• secure computation servers can be set up and operated by independent entities.
• the method and system may be based on software and commodity hardware and therefore may be less expensive to build and operate than any solution involving specialized hardware.
• a cryptographic operation can be applied to the mobile agent comprising the request or a result.
• integrity for the mobile agent can be advantageously guaranteed.
• the originator request is formed by applying an encrypted circuit construction, then the advantage occurs that a secure computation of the request can be achieved and the originator of that request can define how much information an entity shall receive.
• the originator request may comprise a function in encrypted form. This is advantageous, because then other entities processing the originator request can not derive useful information therefrom except results which are designated to those entities.
• the originator request, the first-modified request, the computation-entity result, and the first-entity result may comprise an encrypted part. This is advantageous, because then sensitive information is protected and is not readable as plaintext.
• the originator request may comprise an offer or any other legal instrument. It may also comprises purchase information or customer information, such as the address for delivering or the credit card number for financial transactions.
• the first entity comprises a web server offering a service or goods.
• This service can merely be everything including sale, lease, license, or financing transaction.
• the first-entity result may comprise a customer information concerning acceptance of the originator request. This is advantageous, because then the first-entity can deliver its service or goods immediately and initiate the necessary transactions.
• the customer result may comprise the first-entity information concerning acceptance of the originator request.
• the customer therefore knows that its originator request will be fulfilled and no further actions are necessary.
• FIG. 1 shows an illustration of a communication flow according to the present invention.
• FIG. 2 shows a more detailed illustration of a communication flow.
• FIG. 3 shows another illustration of a communication flow.
• a defining element of a mobile code or agent computation is that it proceeds autonomously and independently of the originator of the code.
• the secure mobile agent computation is modeled in principle as depicted in Fig. 1, whereby the boxes are labeled according to the description below.
• Fig. 1 shows a network, such as the Internet, whereby the code originator O is connected to a first entity H h that is further connected to a second entity H 2 .
• the second entity H 2 is connected to an entity labeled with H j that further has a connection to an entity labeled with H j+ ⁇ , whereby several other entities are possible in-between as indicated by the dotted line.
• the entity H 7+ ⁇ is connected to an entity labeled with H t that in the following is connected back to the code originator O.
• Each of the entities Hi, H 2 , .... Hj, Hj + _, Hi is connected to a computing entity T.
• the code originator O as well as each entity H H 2. .... Hj, H, + ⁇ , Hi sends and receives only a single message, that comprises the agent.
• Fig. 2 depicts a more detailed illustration of the communication flow as shown in Fig. 1.
• the code originator O that here is a customer O
• H the first entity
• the method for processing an originator request OR e.g. a price request including a threshold for a product, of the customer O runs as follows.
• the customer O sends the originator request OR to the first entity H. This is indicated by the arrow labeled with m 0 R.
• the first entity H offers several products to a particular price.
• the first entity H connects to the computation entity T and adds, on receipt of the originator request OR, information I concerning the originator request OR, e.g.
• This first-modified request FMR is sent to the computation entity T as indicated by the arrow labeled with m FM R- If the computation entity T has received the first-modified request FMR it derives therefrom a computation-entity result CER without learning anything from this computation. Afterwards, the computation-entity result CER is sent back to the first entity H as indicated by the arrow labeled with m C ER- If the first entity H has received the computation-entity result CER it derives therefrom a first-entity result EER and forwards this back to the customer O as indicated by the arrow labeled with m FE R.
• the customer O is able to derive from the first-entity result EER a customer result CR.
• This customer result CR provides the information to the customer O whether or not his originator request OR has been fulfilled.
• the first entity H knows form the first-entity result EER whether or not its offer is acceptable to the customer O.
• the originator request OR may contain information about the customer O, i.e. address, credit card information, which allows the first entity H to deliver the requested product immediately.
• the mobile agent visits several vendor sites and compares offers.
• the originator request OR does not only be based on price, but can also include other attributes.
• the originator or customer O wants to maintain the privacy of his preferences, but a shop has an interest to learn the buyer's strategy as well as information about other vendor's offers.
• the vendor wants to keep its method of calculating the price secret. All these requirements can be fulfilled by the described scheme for secure mobile code.
• a shopping agents is traveling around the network and collects offers from several vendors or entities Hi, H 2 , ..., H j , H, + ⁇ , Hi, whereby a prior agreement on the data format of the offers is helpful.
• Fig. 3 shows another embodiment using the same or like parts as shown in Figs. 1 and 2. The difference is the code originator 0 sends directly to each entity H ; , H 2 , ..., H, its originator request OR, whereby the originator request OR can be the same or different to each entity Hi,
• an electronic negotiation is described.
• the electronic negotiation between a buyer and a single vendor can take place using the scheme for secure mobile code that visits a single host or entity H.
• the vendor would act as the originator O and download an applet to the buyer's browser (as is already quite common on the Internet).
• the applet is executed using the help of the computation entity T by the buyer and the offer is displayed to the buyer.
• the vendor may obtain some information as well, which it would have to spell out clearly in a "privacy statement" accompanying the applet.
• Bidding agents can implement a complex strategy being a function of time and other participants' behavior, which gives the bidder more flexibility compared to traditional single-parameter auctions based purely on price.
• a bidder is interested to define his bidding behavior as dynamically as possible, for example making the valuation of a lot depend on other winning bids that he observed in the previous rounds. If the bidders can express their strategies as a computable function, then one may construct a circuit to compute the auction function, i.e., the outcome of the auction, with the strategies as the private inputs of all participants. This would require an auction agent that visits each bidder only once.
• each round of the auction could also be performed securely by an auction applet that visits each bidder once and returns to the auctioneer. There it outputs the winning bids or the end of the auction if the bids did not exceed the minimum increment. If the scheme for secure mobile computing is used, then there is no single entity Hi, H 2 , ..., H j , H j+ ⁇ , Hi that sees all bids (like the auctioneer, its computer system, or its operators). Generalized auctions are common in electricity markets, equities trading, bandwidth auctions, and transportation exchanges, and bidders often have preferences over combination of items.
• the above scheme assumes that the order in which the agent visits all entities is fixed. It is extended to allow for the sequence to depend on Zj by introducing a function ⁇ : Z / ⁇ 1, ..., ⁇ and sending the agent to H ⁇ ( z; ) from H j .
• the function g yields O's output ⁇ and h gives H's oulput z.
• a computation entity T that can be a generic secure computation service, is provided.
• This computation entity T is on-line and connected to all entities Hi, ..., or hosts running agent applications and is at their disposal for securing agent computations.
• the computation entity T itself does not gain any information about the computation, no matter how it behaves, under the assumptions that (1) the computation entity T does not collude with the originator against any entity, and (2) the computation entity T does not collude with any entity against the originator or against any other entity. All computations proceed with minimal or no interaction.
• the scheme is generic and not bound to any particular application. Hence the service of the computation entity T might be offered as a public service for "secure mobile agent computation" on the Internet.
• the scheme is based on functions in encrypted form. For example, encrypting a binary digital circuit realizes the part of the agent computation. This can be realized by an encrypted circuit construction as described in the following.
• G is encrypted
• JJ Alice and Bob engage in a protocol for oblivious transfer, for example, as disclosed by S. Even, O. Goldreich and A. Lempel in "A randomized protocol for signing contract", Communications of the ACM 28 (1985), 637-647 or by G. Brassard, C. Crepeau, and J.-M. Robert in "Information-theoretic reductions among disclosure problems", Proc. 27th IEEE Symposium on Foundations of Computer Science (FOCS), 1986.
• This is an interactive two-party protocol for a sender with input two messages mo and m_ and a chooser with input a bit ⁇ . At the end, the chooser receives m ⁇ but does not learn anything about rn ⁇ ® ⁇ , and the sender has no information about ⁇ .
• the second algorithm ..., K' ⁇ y takes as inputs the encrypted circuit, a representation of x, and a representation of y by the respective keys. It outputs the keys U , ..., U réelle' z from which Bob can recover z, and if Alice and Bob obey the protocol, then z
• Block ciphers are very fast cryptographic primitives, even if implemented in software.
• the following describes how to use the encrypted circuit construction for realizing secure mobile code computation with a single or first entity.
• the extension to multiple entities is considered after that.
• the computation entity T publishes the public key of an encryption scheme.
• the corresponding encryption and decryption operations are denoted by E ⁇ (-) and D (-), respectively. All entities can communicate over secure authenticated links, which could be realized by using standard public-key encryption and digital signatures.
• the basic scheme is that O constructs an encrypted circuit G computing the two values ⁇ and z.
• the code originator O sends G to the first entity H, but encrypts all keys in X for T and does not include the key pairs in U which correspond to ⁇ (denoted by x ) so that the first entity H will not learn anything about ⁇ .
• the first entity H selects from X the encrypted keys representing y and invokes the computation entity T to decrypt them in a single round of interaction.
• the first entity H evaluates the circuit and obtains z; it also returns the keys in the circuit output representing ⁇ to O, who can determine ⁇ from this.
• the scheme proceeds in five steps 1) to 5), as described in the following.
• O chooses a string id that uniquely identifies the computation, e.g., containing the name of O, a description of g and h, and a sequence counter.
• O invokes construct(C) and obtains (G, £, X, U) as above with U consisting of n x + n z key pairs in total.
• U x denotes the pairs in It with indices 1, ..., n x and O_ denotes those with indices n x + 1, ..., n x + n z .
• K ⁇ , b E T (id ⁇ i ⁇ Ki, b ).
• the computation entity T should use a public-key cryptosystem that is secure against adaptive chosen-ciphertext attacks, which means non-malleable, as described by D. Dolev, C. Dwork, and M. Naor in "Non-malleable cryptography", SIAM Journal on Computing 30 (2000), no. 2, 391-437.
• the code originator O and the first entity H should also commit to their inputs. In a practical system, all of these can be realized in the so-called "random oracle model" as described, for example, by M. Bellare and P.
• the code originator O prepares one encrypted circuit G for each entity Hi, ..., Hi and incorporates the encrypted state x s . ⁇ from ⁇ Hnto C® fo ⁇ j > 1. This is achieved by using the output keys U ⁇ , ..., Un ⁇ 'ftom C ⁇ -1) for decrypting a hidden representation of the inputs to
• the cryptosystem includes sufficient redundancy such that given a potential key U and a ciphertext c one can determine with high probability whether c results from an encryption under U.
• the modifications to the scheme are as follows.
• Each entity interprets each £/,• as a symmetric key to E, determines which one of the ciphertexts V® 0 and V® t it decrypts, and then decrypts the one that matches. This yields K. , an oblivious representation of the tth bit in the current state xj of the agent. Those keys are then used to evaluate C .
• Hj When Hj has obtained its output from evaluating C®, it forwards all data that it has received from Hj . together with if , ..., Un x to H y + i . At the end of the circle, H returns if to the code originator O.
• the computation entity T generates the code
• the computation entity T In a variation where the roles of O and T are switched, the computation entity T generates the encrypted circuit. Because it is trusted to follow the protocol one does not have to add a costly zero-knowledge proof for correctness of the whole circuit. Therefore, the operations of the other entities and the corresponding proofs ensuring robustness become much simpler.
• the computation entity T has to know g and h for constructing the circuit, but it may obtain a description of C from O in a first protocol message.
• a three-party oblivious transfer protocol is used, as introduced by M. Naor, B. Pinkas, and R. Sumner in "Privacy preserving auctions and mechanism design", Proc. 1st ACM Conference on Electronic Commerce, 1999, in which the role of the chooser is separated among the chooser and a third party, called the receiver.
• the receiver gets the output message ⁇ specified by the chooser, who itself learns nothing. This so-called "proxy" oblivious transfer can be realized using three message flows: from chooser to receiver and from receiver to sender and back.
• the protocol uses also a one-round implementation of standard oblivious transfer between two parties, which can be realized using the method of C. Cachin, J. Camenisch, J. Kilian, and J. M ⁇ ller, published in their article "One-round secure computation and secure autonomous mobile agents", at Proc. 27th International Colloquium on Automata, Languages and Programming (ICALP) (U. Montanari, J. P. Rolim, and E. Welzl, eds.), Lecture Notes in Computer Science, vol. 1853, Springer, July 2000, pp. 512-523.
• ICALP International Colloquium on Automata, Languages and Programming
• H also prepares its input to n y parallel one-round oblivious transfers (playing the role of the chooser), one for each bit of y. It sends these to T, together with the descriptions of C and Eo(-)', T will send the key pairs K in the one-round oblivious transfers.
• T invokes construct(Q to obtain 6 and the key pairs £ , X , and U . It replies to H withEoC ⁇ ), G, U z and the final flows in all oblivious transfer protocols.
• H can determine the keys L , ...,L n ' ⁇ representing x and the keys K , ...,L n ' representing y. It runs ev ⁇ luate( (7 , L , ...,L embrace ! x ,K , ...,K' ny ) to obtain U , ..., U' ll ⁇ +nz as above. Then it determines its output z from U n ' ⁇ +1 , ..., U réelle' x+nz and from U z , and it forwards U ⁇ , ..., U n ' ⁇ together with E 0 (U x )to O. This enables O to obtain its output ⁇ .
• the protocol shows an extension of the protocol from a single host or first entity Hi to / hosts Hi, ..., Hi.
• the protocol starts as before for the first host.
• the steps for H 2 , ..., Hi are slightly different: three-party oblivious transfer and encryption under Eo are not used.
• T encrypts the keys ® in the input of C ⁇ and representing the state x j - ⁇ of the agent under the output keys in li ⁇ > from G ⁇ as before V®.
• the keys lt° '_1) can be stored by T between step j - 1 and step j or they can be sent along with the protocol flow and are transmitted to T via Hj .
• i and Hj whereby they are encrypted byEr(-)- h addition, the last host obtains U x encrypted with Eo(-) from T and forwards this to O as above.
• the communication pattern is the same as in the basic scheme: there is one message from O to Hi, one from each Hj . i to Hj and one from to O, plus one communication flow between each host and the computation entity T.
• Robustness can be added by using non-malleable public-key encryption schemes and non-interactive zero-knowledge proofs. However, the result will be much more practical because zero-knowledge proofs are not needed for the potentially large encrypted circuit.
• the encrypted circuit construction can be implemented by a block cipher instead of public-key operations.
• the present invention can be realized in hardware, software, or a combination of hardware and software. Any kind of computer system - or other apparatus adapted for carrying out the method described herein - is suited.
• a typical combination of hardware and software could be a general purpose computer system with a computer program that, when being loaded and executed, controls the computer system such that it carries out the methods described herein.
• the present invention can also be embedded in a computer program product, which comprises all the features enabling the implementation of the methods described herein, and which - when loaded in a computer system - is able to carry out these methods.
• Computer program means or computer program in the present context mean any expression, in any language, code or notation, of a set of instructions intended to cause a system having an information processing capability to perform a particular function either directly or after either or both of the following a) conversion to another language, code or notation; b) reproduction in a different material form.

Landscapes

• Engineering & Computer Science (AREA)
• Software Systems (AREA)
• Theoretical Computer Science (AREA)
• Physics & Mathematics (AREA)
• General Engineering & Computer Science (AREA)
• General Physics & Mathematics (AREA)
• Storage Device Security (AREA)
• Management, Administration, Business Operations System, And Electronic Commerce (AREA)
• Computer And Data Communications (AREA)
• Mobile Radio Communication Systems (AREA)

Priority Applications (1)

Applications Claiming Priority (4)

Publications (1)

Family

ID=8170289

Family Applications (1)

Country Status (8)

Families Citing this family (3)

Family Cites Families (1)

• 2001
  • 2001-10-24 CN CNA01818247XA patent/CN1478222A/zh active Pending
  • 2001-10-24 AU AU2002210814A patent/AU2002210814A1/en not_active Abandoned
  • 2001-10-24 CA CA002426794A patent/CA2426794C/en not_active Expired - Lifetime
  • 2001-10-24 EP EP01978722A patent/EP1368721A2/en not_active Withdrawn
  • 2001-10-24 WO PCT/IB2001/001988 patent/WO2002037242A2/en active IP Right Grant
  • 2001-10-24 KR KR1020037006030A patent/KR100582393B1/ko not_active IP Right Cessation
  • 2001-10-24 IL IL15539401A patent/IL155394A0/xx unknown
  • 2001-10-24 JP JP2002539930A patent/JP4336105B2/ja not_active Expired - Fee Related

Non-Patent Citations (1)

Also Published As

Similar Documents

Legal Events