Publication number | WO2001097443 A2 |

Publication type | Application |

Application number | PCT/US2001/018878 |

Publication date | 20 Dec 2001 |

Filing date | 12 Jun 2001 |

Priority date | 12 Jun 2000 |

Also published as | US20020039420, WO2001097442A2, WO2001097442A3, WO2001097443A3 |

Publication number | PCT/2001/18878, PCT/US/1/018878, PCT/US/1/18878, PCT/US/2001/018878, PCT/US/2001/18878, PCT/US1/018878, PCT/US1/18878, PCT/US1018878, PCT/US118878, PCT/US2001/018878, PCT/US2001/18878, PCT/US2001018878, PCT/US200118878, WO 0197443 A2, WO 0197443A2, WO 2001/097443 A2, WO 2001097443 A2, WO 2001097443A2, WO-A2-0197443, WO-A2-2001097443, WO0197443 A2, WO0197443A2, WO2001/097443A2, WO2001097443 A2, WO2001097443A2 |

Inventors | Hovav Shacham, Dan Boneh, Sanjay Beri |

Applicant | Ingrian Systems, Inc. |

Export Citation | BiBTeX, EndNote, RefMan |

Patent Citations (1), Non-Patent Citations (2), Classifications (13), Legal Events (6) | |

External Links: Patentscope, Espacenet | |

WO 2001097443 A2

Abstract

Presented is a method and system for improving the efficiency of network security protections communication protocols such as Secure Socket Layer ('SSL') using enhanced Rivest-Shamir-Adleman ('RSA') encryption and decryption techniques. During the establishment of the initial handshake of SSL communications, where a client is coupled to a server, the server generates a RSA public/private key pair. The public key is formed using two distinct prime numbers. By reducing the size of these prime numbers and arriving at the decrypted message using the Chinese Remainder Theorem, the efficiency of establishing a secure communications session is increased. Likewise if during generation of the public key, the prime numbers possess a mathematical relationship to the public key such that the prime numbers on the order of a third of the size of the public key then the efficiency of establishing the initial handshake is again improved.

Claims (OCR text may contain errors)

What is claimed is:

1. A method for secure computer communications, comprising: generating a Rivest-Shamir-Adleman ("RSA") algorithm public / private key pair at a web server, wherein <N, e'>, represents the public key with N being the product of two distinct primes, p and q, and wherein the private key is represented by d; sending a client hello message to the web server from a client requesting a secure network connection; responding to the client with a server hello message comprising the RSA public key; encrypting a random string R at the client using the RSA public key, wherein the resulting cipher-text C includes R; sending the encrypted cipher-text to the web server; decrypting the cipher-text at the web server using the RSA private key wherein d=rιmod(p-l) and d=r_{2}mod( -l), and wherein <r; , r_{2}> are relatively small numbers on the order of 160 bits in length, wherein R't equals the cipher- text raised to the r; power moduli one ofthe distinct prime numbers and R'_{2} equals the cipher-text raised to the r_{2} power moduli the remaining prime number; combining R'ι and R'_{2} to produce R using the Chinese Remainder Theorem wherein finding R and R'_{2} is more efficient than using standard RSA keys; and establishing a common session key between the web server and client using R.

2. The method of claim 1 , wherein the secure communications includes Secure Socket Layer ("SSL") messages.

3. The method of claim 1 , wherein the secure communications includes Transport Layer Security ("TLS") messages.

4. The method of claim 1 , wherein the secure communications includes internet protocol secure ("IPSec") messages.

5. The method of claim 1, wherein generating a RSA public / private key pair includes; taking the product ofthe n-bit primes to produce an arbitrary number N; picking two random Ar-bit values ri and r_{2} such that n and r2 are on the order of 160 bits and are mathematically related to the n-bit primes and e' is related to N; and sending the public key to a certificate authority and receiving back from the certificate authority a public key certificate for a public key wherein e' is on the order of N in size.

6. The method of claim 5, wherein the k-bit values are related to the n-bit primes by the equations gcd(rι,p - 1) = 1, gcd(r_{2}, q - 1) = 1, and ri = r_{2} mod w, respectively, wherein gcd represents the greatest common divisor and w = gcd(p-l,q-l).

1. The method of claim 6, wherein the relationship between e' and N is expressed by the equation e' = d/^{1} mod φ(N).

8. The method of claim 1, wherein decrypting includes: computing Ri ' and R_{2}' as expressed by the relationship R_{j} = C^{1} mod p and R_{2} = C^{2} modq ; and applying the Chinese Remainder Theorem to produce R, wherein R = R[ modp and R = Rj modq

9. A method for performing an initial handshake during secure communications in a computer network comprising:
coupling a client to a web server; generating a Rivest-Shamir-Adleman ("RSA") algorithm public / private key pair at the web server, wherein the RSA public key is a product of two distinct prime numbers and the private key is a function of two random numbers, wherein each random number has a number of bits greater than or equal to 160 bits and less than a number of bits ofthe RSA key; sending a client hello message to the web server requesting a secure network connection; responding to the client with a server hello message containing the RSA public key; encrypting a random string R at the client using the RSA public key, wherein the resulting cipher-text C includes R; sending the encrypted cipher-text message to the web server; separating cipher-text moduli ofthe two distinct prime numbers; decrypting the moduli of the two distinct prime numbers individually using the two random numbers, wherein the results are combined using the Chinese Remainder Theorem, wherein computational efficiency is improved; and establishing a common session key between the web server and the client using R.

10. The method of claim 9, wherein the initial handshake of secure communications includes Secure Socket Layer ("SSL") messages.

11. The method of claim 9, wherein the initial handshake of secure communications includes Transport Layer Security ("TLS") messages.

12. The method of claim 9, wherein the initial handshake of secure communications includes internet protocol secure ("IPSec") messages.

13. The method of claim 9, further comprising : combining individually encrypted messages into a set of encrypted
messages wherein each encrypted message possesses a public key comprising an encryption exponent; determining a root node of a binary tree containing leaf nodes corresponding to each encryption exponent using a plurality of separate parallel batch trees, wherein the root node of each tree is found and combined to determine the final answer; mimmizing a disparity between sizes ofthe encryption exponents ofthe within the set; using simultaneous multiple exponentiation such that the encryption exponents are combined to reduce the number of exponentiations; calculating a product ofthe encrypted messages; extracting at least one root from the product ofthe encrypted messages; and decrypting the encrypted messages by expressing the at least one root as at least one promise and evaluating the at least one promise at the leaf nodes, and multiplying an inversion of a total product ofthe leaf nodes with a partial product ofthe leaf nodes forming an inversion ofthe leaf node, producing a reduced number of modular inversions wherein efficiency ofthe decryption is increased.

14. The method of claim 9, further comprising keeping the size of N constant while reducing the size ofthe two distinct prime numbers such that the size ofthe two distinct prime numbers is on the order of one third of the size of

N.

15. A method for secure communications, comprising : generating a Rivest-Shamir-Adleman ("RSA") algorithm public / private key pair at a web server, wherein the RSA public key is a product of two distinct prime numbers and the private key is a function of two random numbers; receiving a client hello message from a client requesting a secure socket layer ("SSL") coupling;
responding to the client with a server hello message containing the RSA public key; encrypting a random string R at the client using the RSA public key, wherein the resulting cipher-text includes R; receiving the encrypted cipher-text message at the web server; separating cipher-text moduli ofthe two distinct prime numbers; decrypting the moduli ofthe distinct prime numbers individually using the two random numbers, wherein the results are combined using the Chinese Remainder Theorem; and establishing a common session key between the web server and client using R.

16. A method for secure computer communications, comprising: coupling a web server to a client wherein the client requests the formation of a secure network connection; generating a Rivest-Shamir-Adleman("RSA") algorithm public / private key pair, the public key comprising a root N, wherein N ofthe RSA public key is the product of two distinct n-bit prime numbers, p and q, wherein an encryption exponent e' ofthe RSA public key is ofthe same order in size as the public key root, N encrypting a plain-text message R using the RSA public key such that the resulting text is cipher-text C; decrypting the cipher-text C using the RSA private key wherein the RSA private key is a function of two roots r; and r_{2}, wherein the two roots each are on the order of 160 bits in length; and using the plain-text message R to determine a session encryption key and a session integrity key.

17. A method for Rivest-Shamir-Adleman ("RSA") decryption of secure network communications, comprising: generating a RSA public/private key pair at a web server, wherein <N, e> represents the public key that is mathematically related to two distinct prime
numbers; keeping a size of N constant while reducing a size ofthe two distinct prime numbers by calculating N from a product of a first distinct prime number raised to the first power and a second distinct prime number wherein the first power is greater than one; using the public key by a client to encrypt a plain-text message R to form a cipher-text message C; decrypting the cipher-text C at the web server by using the RSA private key d to determine the plain-text message R by finding R'ι and R'_{2}, wherein the private key is a function of two random numbers <r_{7} , r_{2}>, and wherein an additional R' is constructed by using one ofthe two distinct prime numbers raised to a power greater than one, wherein efficiency ofthe decryption is increased in response to the reduced size ofthe two distinct prime numbers; and computing the plain-text message using the Chinese Remainder Theorem.

18. The method of claim 17, further comprising; combining individually encrypted network security protection handshake messages into a set of encrypted messages wherein each encrypted message is derived using a public key containing an encryption exponent; determining a root node of a binary tree comprising leaf nodes corresponding to each encrypted messages encryption exponent by using a plurality of separate, parallel batch trees finding the root node of each tree and combining the final answers; minimizing the disparity between the sizes of the encryption exponents ofthe public keys within the set; using simultaneous multiple exponentiation such that the encryption exponents are combined to reduce the number of exponentiations; calculating a product ofthe encrypted messages; extracting at least one root from the product ofthe encrypted messages; and decrypting the encrypted messages by expressing the at least one root as
at least one promise and evaluating the at least one promise at the leaf nodes, and multiplying an inversion of a total product ofthe leaf nodes with a partial product ofthe leaf nodes forming an inversion ofthe leaf node wherein the decryption is increased by reducing the number of modular inversions.

19. The method of claim 17, wherein the k-bit values ri, r_{2} are related to the n-bit primes by the greatest common divisor of (rι,p - 1) = 1, (r_{2}, q- l) = l, n = r mod w respectively such that d= r_{\}, modp - l, d= r_{2} mod q - 1, and w is equal to the greatest common divisor of (p-1, q-1).

20. The method of claim 17, wherein decrypting includes: computing R , R"ι, and R_{2}'as expressed by the relationships

21. A method for generating a Rivest-Shamir-Adleman ("RSA") public / private key pair in secure network couplings, comprising: generating two n-bit distinct prime numbers; computing a public key root from a mathematical relationship between two distinct prime numbers; reducing the size ofthe two distinct prime numbers while keeping the size ofthe public key root constant using exponentiation ofthe two distinct prime numbers; forming a public RSA key pair by associating the public key root and a standard RSA encryption exponent; and computing a private RSA key pair by mathematically combining the standard RSA encryption exponent and the n-bit distinct prime numbers.

22. The method of claim 21 , wherein computing the public key root includes the product ofthe square of one n-bit prime number and the second n- bit prime number.

23. The method of claim 21 , wherein the public RSA key pair is indistinguishable from a standard RSA pair.

24. The method of claim 21, further comprising: encrypting a pre-master-secret using the public RSA key pair; and decrypting the pre-master-secret using the private RSA key pair wherein Hensle lifting compensates for reducing the size ofthe distinct prime numbers.

25. A method for Rivest-Shamir-Adleman ("RSA") decryption of secure network communications, comprising: generating a RSA public/private key pair at a web server, wherein <N, e> represents a public key that is mathematically related to two distinct prime numbers and d represents a private key that is mathematically related to two random numbers; keeping a size of N constant while reducing a size ofthe two distinct prime number by calculating N from a product of a first distinct prime number raised to a power greater than one and the second distinct prime number; using the public key at a client to encrypt a plain-text message R to form a cipher-text message C; decrypting the cipher-text C at the web server using the RSA private key d to determine the plain-text message R by finding R'ι and R'_{2}, wherein an additional R"ι is constructed by raising the first ofthe two distinct prime numbers to a power greater than one, wherein the efficiency ofthe decryption is increased due to a reduced size ofthe two distinct prime numbers using the private RSA key pair, wherein Hensle lifting compensates for altering a multiplicity ofthe distinct prime numbers; and computing the plain-text message using the Chinese Remainder Theorem.

26. A method for Rivest-Shamir-Adleman ("RSA") decryption of secure network communications, comprising: generating a RSA public/private key pair at the web server wherein <N,
e> represents the public key that is mathematically related to two distinct prime numbers; keeping a size of N constant while reducing a size ofthe two distinct prime numbers such that each ofthe two distinct prime numbers is on the order of one third ofthe size of N; using the public key at a client to encrypt a plain-text message R to form a cipher-text message C; decrypting the cipher-text C at the web server by using the RSA private key d, to determine the plain-text message R by finding R'ι and R'_{2}, wherein an additional R"ι is constructed by using the one ofthe two distinct prime numbers raised to a power greater than one, wherein the efficiency ofthe decryption is increased in response to the reduced size ofthe two distinct prime numbers using the private RSA key pair wherein Hensle lifting compensates for altering the multiplicity ofthe distinct prime numbers; and computing the plain-text message using the Chinese Remainder

Theorem.

27. A system for Rivest-Shamir-Adleman ("RSA") decryption of secure network communications, comprising: a web server generating a RSA public/private key pair wherein <N, e> represents a public key that is mathematically related to two distinct prime numbers; the web server keeping a size of N constant while reducing a size ofthe two distinct prime numbers by calculating N from the product of a first distinct prime number raised to a power greater than one and a second distinct prime number; a client using the public key to encrypt a plain-text message R to form a cipher-text message C; the web server decrypting the cipher-text C by using the RSA private key d to determine the plain-text message R by finding R and R^{f} _{2}, wherein an additional R"ι is constructed by using one ofthe two distinct prime numbers raised to a power greater than one wherein the efficiency ofthe decryption is
increased in response to the reduced size ofthe two distinct prime numbers; and the web server computing the plain-text message using the Chinese Remainder Theorem.

28. A system for using Rivest-Shamir-Adleman ("RSA") decryption of secure network communications in a computer network, comprising: at least one web server; at least one client processor coupled to the at least one web server, wherein the at least one web server generates a RSA public/private key pair, <N, e>, representing the public key that is mathematically related to two distinct prime numbers, wherein d represents the private key; the at least one web server keeping a size of N constant while reducing a size ofthe two distinct prime numbers by calculating N from the product of a first distinct prime number raised to a power greater than one and a second distinct prime number; the at least one client processor using the public key to encrypt a plaintext message R to form a cipher-text message C; the at least one web server decrypting the cipher-text message C by using the RSA private key <π , r_{2}> to determine the plain-text message R by finding R'ι and R'_{2}, wherein an additional R"ι is constructed by using one ofthe two distinct prime numbers raised to a power greater than one wherein the efficiency ofthe decryption is increased in response to the reduced size ofthe two distinct prime numbers; and the at least one web server computing the plain-text message using the Chinese Remainder Theorem.

29. A computer-readable medium, comprising executable instructions for Rivest-Shamir-Adleman ("RSA") decryption of secure network communications which, when executed in a processing system, causes the system to: couple a web server to a client; send a client hello message to the web server requesting a secure
network connection; generate a Rivest-Shamir-Adleman ("RSA") algorithm public / private key pair at the web server wherein the RSA public key is the product of two distinct prime numbers wherein the RSA private key is a function of two random numbers wherein each random number has a number of bits greater than or equal to 160 bits and less than a number of bits ofthe RSA key; respond to the client with a server hello message containing the RSA public key; encrypt a random string R at the client using the RSA public key, wherein the resulting cipher-text C includes R; send the encrypted cipher-text message C to the web server; separate cipher-text C moduli ofthe two distinct prime numbers; decrypt the moduli ofthe two distinct prime numbers individually using the two random numbers, wherein results are combined using the Chinese Remainder Theorem, wherein computational efficiency is improved and establish a common session key between the web server and the client using R.

30. An electromagnetic medium, comprising executable instructions for Rivest-Shamir-Adleman ("RSA") decryption of secure network communications which, when executed in a processing system, causes the system to: couple a web server to a client; send a client hello message to the web server requesting a secure network connection; generate a Rivest-Shamir-Adleman ("RSA") algorithm public / private key pair at the web server wherein the RSA public key is the product of two distinct prime numbers, wherein the RSA private key is a function of two random numbers wherein each random number has a number of bits greater than or equal to 160 bits and less than a number of bits of the RSA key; respond to the client with a server hello message containing the RSA public key;
encrypt a random string R at the client using the RSA public key, wherein the resulting cipher-text C includes R; send the encrypted cipher-text message C to the web server; separate cipher-text moduli ofthe two distinct prime numbers; decrypt the moduli ofthe two distinct prime numbers individually using the two random numbers, wherein results are combined using the Chinese Remainder Theorem, wherein computational efficiency is improved; and establish a common session key between the web server and the client using R.

31. A computer-readable medium, comprising executable instructions for Rivest-Shamir-Adleman ("RSA") decryption of secure network communications which, when executed in a processing system, causes the system to: generate a RSA public/private key pair at the web server wherein <N, e> represents the public key that is mathematically related to two distinct prime numbers; keep a size of N constant while reducing a size ofthe two distinct prime numbers such that each ofthe two distinct prime numbers is on the order of one third ofthe size of N; use the public key at client to encrypt a plain-text message R to form a cipher-text message C; decrypt the cipher-text C at the web server by using the RSA private key d to determine the plain-text message R by finding R and R'_{2}, wherein an additional R" i is constructed by using one ofthe two distinct prime numbers raised to a power greater than one, wherein the efficiency ofthe decryption is increased in response to the reduced size ofthe two distinct prime numbers using the private RSA key pair wherein Hensle lifting compensates for altering the multiplicity ofthe distinct prime numbers; and compute the plain-text message using the Chinese Remainder Theorem.

32. An electromagnetic medium, comprising executable instructions for Rivest-Shamir-Adleman ("RSA") decryption of secure network communications which, when executed in a processing system, causes the system to: generate a RSA public/private key pair at the web server wherein <N, e> represents the public key that is mathematically related to two distinct prime numbers; keep a size of N constant while reducing a size ofthe two distinct prime numbers such that each ofthe two distinct prime numbers is on the order of one third of the size of N; use the public key at a client to encrypt a plain-text message R to form a cipher-text message C; decrypt the cipher-text C at the web server by using the RSA private key d to determine the plain-text message R by finding R'i and R'_{2}, wherein an additional R" _{1} is constructed by using one ofthe two distinct prime numbers raised to a power greater than one, wherein the efficiency ofthe decryption is increased in response to the reduced size ofthe two distinct prime numbers using the private RSA key pair wherein Hensle lifting compensates for altering the multiplicity ofthe distinct prime numbers; and compute the plain-text message using the Chinese Remainder Theorem.

Description (OCR text may contain errors)

METHOD AND APPARATUS FOR ENHANCING NETWORK SECURITY PROTECTION SERVER PERFORMANCE

FIELD OF THE INVENTION The claimed invention relates to the field of secure communications.

More particularly it relates to improving the efficiency of secure network communications.

BACKGROUND OF THE INVENTION Many network transactions today require secure communications. To establish a secure communication link protocols such as Secure Socket Layer ("SSL") and Transport Layer Security ("TLS") must be accomplished. Today SSL is the most widely deployed protocol for securing communication on the World Wide Web ("WWW"). The protocol is used by most E-commerce and financial web sites as it guarantees privacy and authenticity of information exchanged between a web server and a web browser. Currently, the number of web sites using SSL to secure web traffic is growing at a phenomenal rate and as the services provided on the World Wide Web continue to expand so will the need for security using SSL. Unfortunately, neither SSL or TLS are cheap. A number of studies have shown that web servers using the SSL protocol perform far worse than web servers that do not encrypt web traffic. In particular, a web server using SSL can handle 30 to 50 times fewer transactions per second than a web server using clear-text communication only can. The exact transaction performance degradation depends on the type of web server used by the site. To overcome this degradation web sites using secure connections typically buy significantly more hardware in order to provide a reasonable response time to their customers.

Web sites often use one of two techniques to overcome security's impact on performance. The first method, as indicated above, is to deploy more machines at the web site and load balance connections across these machines. This is problematic since more machines are harder to administer. In addition, mean time between failures decreases significantly. The other solution is to install a hardware acceleration card inside the web server. The card handles most ofthe secure protocol workload thus enabling the web server to focus on its regular tasks. Accelerator cards are available from a number of vendors and while these cards reduce the penalty of using secure protocols, they are relatively expensive and are non-trivial to configure. Thus there is a need to quickly establish secure transactions at a lower cost.

SUMMARY OF THE INVENTION A method and apparatus for enhancing security protection server performance in a computer network is provided when a web browser first connects to a web server using secure protocols, the browser and server execute an initial handshake protocol. The outcome of this protocol is a session encryption key and a session integrity key. These keys are only known to the web server and web browser, and establish a secure session.

Once session keys are established, the browser and server begin exchanging data. The data is encrypted using the session encryption key and protected from tampering using the session integrity key. When the browser and server are done exchanging data the connection between them is closed. This process begins when the web browser connects to the web server and sends a client-hello message. Soon after receiving the message, the web server responds with a server-hello message. This message contains the server's public key certificate that informs the client ofthe server's Rivest-Shamir-Adleman algorithm ("RSA") public key. Having received the public key, the browser picks a random 48-byte string, R, and encrypts it using the key. Letting C be the resulting cipher-text ofthe string R, the web browser then sends a client- key-exchange message containing C. The 48-byte string R is called the pre- master- secret. Upon receiving the message, from the browser, the web server uses its RSA private key to decrypt C and thus learns R. Both the browser and server then use R and some other common information to derive the session keys. With the session keys established, encrypted message can be sent between the browser and server with impunity. The decryption ofthe encrypted string, R, is the expensive part ofthe initial handshake. An RSA public key is made of two integers (N, e). In an embodiment N = pq is the product of two large primes and is typically 1024 bits long. The value e is called the encryption exponent and is typically some small number such as e = 65537. Both N and e are embedded in the server's public key certificate. The RSA private key is simply an integer d satisfying e ■ d = 1 mod (p - 1) (q - 1). Given an RSA cipher-text C, the web server decrypts C by using its private key to compute mod N that reveals the plain-text message, R. Since both d and N are large numbers (each 1024 bits long) this computation takes some effort.

At a later time, the browser may reconnect to the same web server. When this happens the browser and server executes the resume handshake protocol. This protocol causes both server and browser to reuse the session keys established during the initial handshake saving invaluable resources. All application data is then encrypted and protected using the previously established session keys.

Ofthe three phases, the initial handshake is often the reason why secure connections degrade web server performance. During this initial handshake the server performs an RSA decryption or an RSA signature generation. Both operations are relatively expensive and the high cost ofthe initial handshake is the main reason for supporting the resume handshake protocol. The resume handshake protocol tries to alleviate the cost ofthe initial handshake by reusing previously negotiated keys across multiple connections. However, in the web environment, where new users constantly connect to the web server, the expensive initial handshake must be executed over and over again at a high frequency. Hence, the need for reducing the cost ofthe initial handshake protocols.

One embodiment presents an implementation of batch RSA in an SSL web server while other embodiments present substantial improvements to the basic batch RSA decryption algorithms. These embodiments show how to reduce the number of inversions in the batch tree to a single inversion. Another embodiment further speeds up the process by proper use ofthe Chinese Remainder Theorem ("CRT") and simultaneous multiple exponentiation. While the Secure Socket Layer ("SSL") protocol is a widely utilized technique for establishing a secure network connection, it should be understood that the present invention can be applied to the establishment of any secure network based connection using a plurality of protocols.

A different embodiment entails architecture for building a batching secure web server. The architecture in this embodiment is based on using a batching server process that functions as a fast decryption oracle for the main web server processes. The batching server process includes a scheduling algorithm to determine which subset of pending requests to batch.

Yet other embodiments improve the performance by reducing the handshake work on the server per connection. One technique supports web browsers that deal with a large encryption exponent in the server's certificate, while another approach supports any browser.

BRIEF DESCRIPTION OF THE DRAWINGS The present invention is illustrated by way of example in the following figures in which like references indicate similar elements. The following figures disclose various embodiments ofthe claimed invention for purposes of illustration only and are not intended to limit the scope ofthe claimed invention.

Figure 1 is a flow diagram ofthe initial handshake between a web server and a client of an embodiment.

Figure 2 is a flow diagram for increasing efficiency ofthe initial handshake process by utilizing cheap keys of an embodiment. Figure 3 is a flow diagram for increasing efficiency ofthe initial encryption handshake by utilizing square keys in an embodiment.

Figure 4 is a block diagram of an embodiment of a network system for improving secure communications.

Figure 5 is a flow diagram for managing multiple certificates using a batching architecture of an embodiment.

Figure 6 is a flow diagram of batching encrypted messages prior to decryption of an embodiment. DETAILED DESCRIPTION

The establishment of a secure network connection can be improved by altering the steps ofthe initial handshake. One embodiment for the improvement to the handshake protocol focuses on how the web server generates its RSA key and how it obtains a certificate for its public key. By altering how the browser uses the server's public key to encrypt a plain-text R, and how the web server uses its private key to decrypt the resulting cipher-text C, this embodiment provides significant improvements to Secure Socket Layer ("SSL") communications. While the Secure Socket Layer protocol is a widely utilized technique for establishing a secure network connection, it should be understood that the techniques described herein can be applied to the establishment of any secure network-based connection using any number of protocols.

The general process in establishing a Secure Socket Layer communication between a browser or client and a server or host is depicted in

Figure 1. The process begins with a request from the browser to establish a secure session 110. The client forms a hello message requesting a public key and transmits the message to the server 114. Upon receiving the client-hello message, the web server responds with a server-hello message containing a public key 118. The public key is one half of a public / private key pair. While the server transmits the public key back to the browser the server keeps the private key. Once the client receives the public key 122 a random number R is generated 126. This random number is the session key. The client encrypts R by using the private key that it received from the server 132. With the number R encrypted, the client sends the cipher-text to the web-server 138. Upon receiving the cipher-text 142 the web server user the private key portion ofthe public / private key pair to decrypt the cipher-text 146. With both the client and the server possessing the session key R, a new encrypted secure socket layer session 160 is established using R as the session key 158. This session is truly encrypted since only the client and the web server possess the session key for encryption and decryption.
In one embodiment that improves the establishment of a secure connection a server generates an RSA public/private key pair by generating two distinct n-bit primes p and q and computing N =pq. While N can be of any arbitrary size, assume for simplicity that N is 1024 bits long and let w - cd(p - l, q - X) where gcd is the greatest common divisor. The server then picks two random Ar-bit values r\, r_{2} such that gcd(n, p - 1) = 1, gcd(r_{2}, q - l) = l, and r_{!} = r_{2} mod w. Typically k falls in the range of 160 -512 bits in size. Although other larger values are also acceptable, k is minimized to enhance performance. The server then computes d such that d = r_{\} odp - 1 and d = r_{2} mod q - 1. Having computed d, e' is found by solving the equation e' = d/^{1} mod φ(N) resulting in the public key being (N, e') and the private key d, which is a function of two random numbers, (ri, r_{2}).

The server then sends the public key to a Certificate Authority (CA). The CA returns a public key certificate for this public key even though e' is very large, namely on the order of N. This is unlike standard RSA public key certificates that use a small value of e, e.g. e = 65537. Consequently, the CA must be willing to generate certificates for such keys.

To find dthe Chinese Remainder Theorem is typically used. Unfortunately, p - 1 and q - 1 are not relatively prime (they are both even) and consequently the theorem does not apply. However, by letting w - gcd(p - \, q

- 1), knowing that — — and - — are relatively prime, and recalling that n = w w r_{2} = a mod w, the CRT can be used to find an element d' such that

Observing that the required d is simply d = w • ά + a and indeed, d = ri mod p - 1 and d = r_{2} mod q - 1 , if w is large, the requirement that ri = r_{2} mod w reduces the entropy ofthe private key. For this reason it is desirable to ensure that w is small and one embodiment lets w = 2, or namely that gcd(p - 1, q - 1) = 2. Recall that gcd(rι, p - 1) = 1 and gcd(r_{2}, q - ϊ) = l.
It follows that gcd(d,p- 1) = 1 and gcd(d, q - 1) = 1 and consequently gcd(d, (p - l)(q - 1)) = 1. Hence, d is invertible modulo φ(N) = (p - l)(q - 1).

The web browser obtains the server's public key certificate from the server-hello message. In this embodiment, the certificate contains the server's public key (N, e). The web browser encrypts the pre-master-secret R using this public key in exactly the same way it encrypts using a normal RSA key. Hence, there is no need to modify any ofthe browser's software. The only issue is that since e' is much larger than e in a normal RSA key, the browser must be willing to accept such public keys. When the web server receives the cipher-text C from the web browser the web server then uses the server's private key, (ri, ri), to decrypt C. To accomplish this the server computes R{ = C^{rχ} odp and R'_{2} = C^{f2} modg. Using CRT the server then computes an R e Z_{Λ}Γ such that R = R{ mod p and R = R_{2} mod q , noting that R = C^{d} mod N. Hence, the resulting R is a proper decryption of C.

Decryption using a standard RSA public key is completed with C^{d} mod Nusing the CRT. Typically Ri = < ^{dmodP}- » modp and R_{2} = ( ^{dmoά 9}- D _{mo}d q is first computed and then the CRT is applied to Ri, R_{2} to obtain R mod N. Note that the exponents dmodp — 1 and mod q — 1 are typically as large as p and q, namely 512 bits each. Hence, to generate the signature the server must compute one exponentiation modulo j? and one exponentiation modulo q. When Nis 1024 bits, the server does two full exponentiations modulo 512-bit numbers.

In one embodiment, the server computes R{ , R_{2} and then applies CRT to R{ , R_{2}. As in normal RSA, the bulk of the work is in computing R[, R_{2}' . However, computing R{ requires raising C to the power of ri, which is minimized. Since the time that modular exponentiation takes is linear in time to the size ofthe exponent, computing R{ takes approximately one third the work and one third ofthe time of raising Cto the power of a 512 bit exponent. Hence, computing R{ takes one third the work of computing Ri. Therefore,
during the entire decryption process the server does approximately one third the work as in a normal SSL handshake.

To illustrate the implementation of this embodiment suppose Eve is an eavesdropper that listens on the network while the handshake protocol is taking place. Eve sees the server's public key (N, e') and the encrypted pre-master- secret C. Suppose ri < r_{2}. It can be shown that an adversary who has (N, e', C) can mount an attack on the system that runs in time

Let (N, e ') be an RSA public key with N =pq and let d eZ e, the corresponding RSA private key satisfying d = rj, modp - 1 and d = r_{2} mod q - 1 with r; < r_{2}. If rt is m bits long and it is assumed that r ≠ r_{2} mod 2^{m 2}, then given (N, e') aa. adversary can expose the private key d in time

then if follows that G(g

{s^{e'}-^{2,}"^{/2}-^{A} -_{8} ^{e'}-^{B} -g) = _{S} ^{e'}" -g - (modp).

Since ri ≠ r_{2} mod 2'"^{/2}, it can be shown that G(g^{e' B}) ≠ 0 mod q. Hence, gcd (N, G(g^{e' ' B})) gives a non-trivial factor of N. Hence, if G(x) mod N is evaluated at x

= g^{e' 'j} for j = 0, . . . , 2^{m 2} at least one ofthe values will expose the factorization of N. Evaluating a polynomial of degree 2^{ml2} at 2^{ml2} values can be done in time 2^{ml2} • m/2 using Fast Fourier Transform methods. This algorithm requires

Thus in order to obtain security of 2 °, both ri and r

Figure 2 is a flow diagram for improving secure socket layer communications of an embodiment by altering the public / private key pair. In operation, the server generates an RSA public / private key pair initiating a normal initial handshake protocol 210. At this point the server generates two distinct prime numbers 215 and takes the product ofthe numbers to produce the N component ofthe public key 220. Similarly, the server picks two random values to create the private key 225. Using the prime numbers, 215, and the random values ofthe private key 225 the server computes the value d, 230, and correspondingly the value e 235. The result is a new public / private key pair

240 that the client uses to encrypt the pre-master-secret R 250. Once R has been encrypted with the new public key and transmitted to the server as cipher- text C, the server uses it private key to decrypt the pre-master-secret 260. Once Ri and R_{2} have been determined 265 they are combined to find R 270. Having the value ofthe pre-master-secret intact, the server and client can establish a secure session 280.

A further embodiment dealing with the handshake protocol reduces the work per connection on the web server by a factor of two. This embodiment works with all existing browsers. As before, the embodiment is illustrated by describing how the web server generates its RSA key and obtains a certificate for its public key. This embodiment continues in describing how the browser uses the server's public key to encrypt a plain-text R, and the server uses its private key to decrypt the resulting cipher-text C.

In this embodiment the server generates an RSA public/private key pair by generating two distinct ra-bit primes p and q such that the size of each distinct prime number is on the order of one third ofthe size of N. Using this relationship the server computes JΨ as N' =p^{2} • q. The relationship between the prime numbers and N is dependent on the power by which one ofthe prime number is raised. For example if one ofthe prime numbers was raised to the fourth power the prime numbers would be on the order of one fifth the size of

N. The exponent of at least one ofthe prime numbers must be greater than one. While clearly JV^{7} can be of arbitrary size, assume, in the situation where p is raised to the power of two and q is raised to the power of one, that N is 1024 bits long, and hence p and q are 341 bits each instead ofthe typical 512 bits. The server uses the same e used in standard RSA public keys, namely e = 65537 as long as gcd(e, (p - 1) (q - ϊ)) = 1. The server then computes d= e^{~l} mod (p — l)(q — 1) as well as ri = d mod - 1 and r_{2} = d mod q - 1. With the public
key being (JV, e) and the private key being d, which is a function of (n, ri), the server sends the public key, (N, e), to a Certificate Authority (CA) and the CA returns a public key certificate. The public key in this case cannot be distinguished from a standard RSA public key. The web browser obtains the server's public key certificate from the server-hello message. The certificate contains the server's public key (N^{7}, e). The web browser encrypts the pre-master-secret R using this public key in exactly the same way it encrypts using a normal RSA key.

When the web server receives the cipher-text C from the web browser the web server decrypts C by computing R{ = C^{'1} mod and R_{2}' = C^{rι} mod q .

Note that (R[ )^{e} = Cmodp and (R'_{2} )^{e} = C mod q. Lifting the server constructs an Rf such that (R")^{e} = C mod ^{2}. More precisely, the server computes

Using CRT, the server computes an R e Z# such that R = Rf mod p and R = R_{2} mod q noting that R = C^{d} mod N. Hence, the resulting R is a proper decryption of C. Recall that when N is 1024 bits, the server typically does two full exponentiations modulo 512-bit numbers. In this embodiment the alteration ofthe multiplicity ofthe roots is compensated by the lifting mechanism.

In this embodiment the server computes R{, R_{2}, R{' and then applies CRT to Rf, R_{2}' . The bulk of the work is in computing R{ , R_{2} , Rf but computing R{ requires a full exponentiation modulo a 341 -bit prime rather than a 512-bit prime. The same holds for R_{2}. Hence in this embodiment, computing R{ , R_{2} takes approximately half the time of computing R_{l5} R_{2}. Furthermore, computing Rf from R{ only requires a modular inversion modulo p^{2}. This takes little time when compared with the exponentiations for computing R{ , R_{2}. Hence, using this embodiment the handshake takes approximately half the work of a normal handshake on the server.
Some accelerator cards do not provide support for modular inversion. As a result, the inversion is preformed using an exponentiation. This is done by observing that for any x Z^{*} _{p} the inverse of x is given by:

_{X}-^{I} _{= X}P^{2}-P-^{I} (modp^{2}). Unfortunately, using an exponentiation to do the inversion hurts performance.

As discussed herein a better embodiment for inversion in this case is batching.

One can invert two numbers x_{x},x_{2} e Z^{*} as a batch faster than inverting the two numbers separately. To do so use the fact that

^{x}ϊ ~ ^{x}2 '(^{x}i^{x}2j^{~ an}& ^{x}2^{l = x}ι ' (^{x}ι^{χ} _{2})^{_1} (modp^{2}). Hence, at the cost of inverting xi ^{•} x_{2} it is possible to invert both i and x_{2}. This embodiment shows that an inversion of £ elements x_{15}...,x_{k} e Z^{*} is at the cost of one inversion and k log_{2} k multiplications. Thus, the amortized cost of a single inversion is Ilk of an exponentiation plus log_{2} k multiplications.

To take advantage of batched inversion in the SSL handshake a number of instances ofthe handshake protocol are collected from among different users and the inversion is preformed on all handshakes as a batch. As a result, the

amortized total number of exponentiations per handshake is 2 + — . This

K approximately gives a factor of two improvement in the handshake work on the server as compared to the normal handshake protocol. The security ofthe improved handshake protocol depends on the difficulty of factoring integers ofthe form N=p^{2 •} q. When 1024 bit keys are used the fastest factoring algorithms (i.e. the number field sieve) cannot take advantage ofthe special structure of N. Similarly, p and q are well beyond the capabilities ofthe Elliptic Curve Method (ECM). Figure 3 is a flow diagram for modifying the public key of an embodiment to facilitate an improvement in secure socket layer communication. As in other embodiments, the process begins with the servers generation of a RSA public / private key pair 310. In this embodiment, the public key is modified. The web server generates two distinct prime numbers 312 and computes a new Ν' 318. Using the same exponent 320 the server computes the
value d 322 which it uses to find the private key 328. The result is a pubic / private key combination 330 that the sever then sends to the client for the encryption ofthe pre-master-secret 340. Once the server receives the encrypted pre-master-secret, R, from the client 350 the server decrypts R 360 by computing RI 362 and R2 368 and combining the results 370. Once R has been determined the client can establish a secure session with the client using the new session key 380.

The establishment of a secure connection between a server and a browser can also be improved by batching the initial SSL handshakes on the web server. In one embodiment the web server waits until it receives b handshake requests from b different clients. It treats these b handshakes as a batch, or set of handshakes, and performs the necessary computations for all b handshakes at once. Results show that, for b = 4, batching the SSL handshakes in this way results in a factor of 2.5 speedup over doing the b handshakes sequentially, without requiring any additional hardware.

One embodiment improves upon a technique developed by Fiat for batch RSA decryption. Fiat suggested that one could decrypt multiple RSA cipher- texts as a batch faster than decrypting them one by one. Unfortunately, experiments show that Fiat's basic algorithm, naively implemented, does not give much improvement for key sizes commonly used in initial secure handshakes.

A batching web server must manage multiple public key certificates. Consequently, a batching web server must employ a scheduling algorithm that assigns certificates to incoming connections, and picks batches from pending requests, so as to optimize server performance.

To encrypt a message using an RSA public key (N, e), the message M is formatted to obtain an integer Xin (1, . . . , N}. This formatting is often done using the PKCS1 standard. The cipher-text is then computed as C = X^{2} mod N. This process occurs during the initial stages ofthe initial handshake between a client and server when attempting to create a secure connection.

To decrypt a cipher-text C the web server uses its private key d o compute the e^{,th} root of C in Z_{Λ}Γ. The e ^{Λ} root of C is given by & mod N as
previously noted. Since both d and N are large numbers (each 1024 bits long) this is a lengthy computation on the web server. It is noted that d must be taken as a large number (i.e., on the order of N) since otherwise the RSA system is insecure. When using small public exponents, ei and e_{2}, which are components of the public key, it is possible to decrypt two cipher-texts for approximately the price of one. Suppose Vi is a cipher-text obtained by encrypting using the public key (N, 3). Similarly, imagine v_{2} is a cipher-text obtained by encrypting using the public key (N, 5). To decrypt v_{\} and v_{2}, computing v and i^{1/5} mod N are made by setting A = (vι^{5 ■} v_{2} ^{3})^{1/15} it can be shown that

Hence, at the cost of computing a single 15

This batching technique is most useful when the public exponents ei and e_{2} are very small (e.g., 3 and 5). Otherwise, the extra arithmetic required can be expensive. Also, only cipher-texts encrypted using distinct public exponents can be batch decrypted. Indeed, it can be shown that it is not possible to batch when the same public key is used. That is, it is not possible to batch the computation of v_{\} ^{β} and v ^{β}. This observation to the decryption of a batch of b RSA cipher-texts can be generalized. In one embodiment there are b distinct and pairwise relatively prime public keys ei, . . . ,e_{δ}, all sharing a common modulus N=pq. Furthermore, assume there are b encrypted messages, vi, . . . , ^, one encrypted with each key, that are desirable to decrypt simultaneously, to obtain the plain- texts m_{{} = v_{t} ^{e}' .

The batch process is implemented around a complete binary tree with b leaves, possessing the additional property that every inner node has two children. In one embodiment the notation is biased towards expressing locally recursive algorithms: Values are percolated up and down the tree. With one exception, quantities subscripted by L or R refer to the corresponding value of
the left or right child ofthe node, respectively. For example, m is the value ofm at a node; m_{R} is the value ofm at that node's right child and so forth.

Certain values necessary to batching depend on the particular placement of keys in the tree and may be pre-computed and reused for multiple batches. Pre-computed values in the batch tree are denoted with capital letters, and values that are computed in a particular decryption are denoted with lower-case letters.

The batching algorithm consists of three phases: an upward-percolation phase, an exponentiation phase, and a downward-percolation phase. In the upward-percolation phase, the individual encrypted messages Vj are combined to form, at the root ofthe batch tree, the value v = π ι ^{e}' . where e = π ?_{=1}e, .

In preparation, assign to each leaf node a public exponent: E - e,-. Each inner node then has its E computed as the product of those of its children: E <— E_ ■ E_{R}. The root node's E will be equal to e, the product of all the public exponents.

Each encrypted message v, is placed (as v) in the leaf node labeled with its corresponding e,-. The v's are percolated up the tree using the following recursive step, applied at each inner node: v <-v£* -v ^{L}. At the completion of the upward-percolation phase, the root node contains v = TT _{=1} v ^{e}' . In the exponentiation phase, the e ^{th} root of this v is extracted. Here, the knowledge of factorization of Nis required. The exponentiation yields v^{Ve} = [ -_{=}ι ^{v})'^{e}' _{>} which is stored as m in the root node. In the downward-percolation phase, the intent is to break up the product m into its constituent subproducts m , and m_{R}, and, eventually, into the decrypted messages m_{t} at the leaves. At each inner node an is chosen satisfying the following simultaneous congruencies:

X= 0 (mod E_{L}) X= 1 (mod E_{R}).

The value X is constructed using the Chinese Remainder Theorem ("CRT"). Two further numbers, Z_{L} and X_{R}, are defined at each node as follows:

X_{L} =X/E_{h} X_{R} = (X- 1)/E_{R}.
Both divisions are done over the integers. (There is a slight infelicity in the naming here: X_{L} andX_{R} are not the same as the s ofthe node's left and right children, as implied by the use ofthe L and R subscripts, but separate values.) The values of X, X_{L}, and X_{R} are such that, at each inner node, m^{x} equals v_{L} ^{L •} v_{R} ^{R •} m_{R} . This immediately suggests the recursive step used in downward-percolation:

At the end ofthe downward-percolation process, each leafs m contains the decryption ofthe v placed there originally. Only one large (full-size) exponentiation is needed, instead of b of them. In addition, the process requires a total of 4 small exponentiations, 2 inversions, and 4 multiplications at each of the b - 1 inner nodes.

Basic batch RSA is fast with very large moduli, but may not provide a significant speed improvement for common sized moduli. This is because batching is essentially a tradeoff. Batching produces more auxiliary operations in exchange for fewer full-strength exponentiations.

Batching in an SSL-enabled web server focuses on key sizes generally employed on the web, e.g., n = 1024 bits. Furthermore, this embodiment also limits the batch size b to small numbers, on the order of b = 4, since collecting large batches can introduce unacceptable delay. For simplicity of analysis and implementation, the values of b are restricted to powers of 2.

Previous schemes perform two divisions at each internal node, for a total of 2b - 2 required modular inversions. Modular inversions are asymptotically faster than large modular exponentiations. In practice, however, modular inversions are costly. Indeed, the first implementation (with b = 4 and a 1024- bit modulus) spends more time doing the inversions than doing the large exponentiation at the root. Two embodiments, when combined, require only a single modular inversion throughout the algorithm with the cost of an additional O(b) modular multiplication. This tradeoff gives a substantial running-time improvement.

The first embodiment is referred to herein as delayed division. An important realization about the downward-percolation phase is that the actual
value of for the internal nodes ofthe tree is consulted only for calculating m_{L} and m_{R}. An alternative representation of that supports the calculation of m_{L} and m_{R} and that can be evaluated at the leaves to yield m would do just as well. This embodiment converts a modular division alb to a "promise," (a, b). This promise can operate as though it were a number, and, can "force" getting its value by actually computing b^{Λ}a. Operations on these promises work in a way similar to operations inprojective coordinates as follows: α/b = ( ,b) ( ,b)^{c} = (α^{c},b^{c}) c -(a,b) = (ac,b) {a,b) -(c,d) = (ac,bd) .

(a, b)lc = (a, be) (a, b) l(c, d) = (ad, be)

Multiplication and exponentiation takes twice as much work had these promises not been utilized, but division can be computed without resort to modular inversion.

If, after the exponentiation at the root, the root m is expressed as a promise, m <— (m, 1), this embodiment can easily convert the downward- percolation step to employ promises: m_{R} <- m /( _{L} ^{L} - _{R} ^{R} ) m_{L} <- m/m_{R} .

No internal inversions are required. The promises can be evaluated at the leaves to yield the decrypted messages.

Batching with promises uses b - 1 additional small exponentiations and b - 1 additional multiplications. This translates to one exponentiation and one multiplication at every inner node, saving 2(b - 1) - b = b - 2 inversions. To further reduce the number of inversions, another embodiment uses batched divisions. When using delayed inversions one division is needed for every leaf ofthe batch tree. In the embodiment using batched divisions, these b divisions can be done at the cost of a single inversion with a few more multiplications. As an example of this embodiment, invert three values x, y, and z.

Continue by forming the partial product^, xz, and xy and then form the total product xyz and invert it, yielding (xyz)^{Λ}. With these values, calculate all the inverses: x^{_1} = (yz) ■ (xyz)^{A} y^{Λ} = (xz) ^{■} (xyz)^{'1} z^{Λ} = (xy) ^{■} (xyz)^{'1}
Thus the inverses of all three numbers are obtained at the cost of only a single modular inverse along with a number of multiplications. More generally, it can be shown that by letting xi, . . . , x„ e Z_{N>} all n inverses x ^{1} , . . . , x^{"1} can be obtained at the cost of one inversion and 3n - 3 multiplications. It can be proven that a general batched-inversion algorithm proceeds in three phases. First, setAι <- i, and A_{t} - x_{t} • Aμ.ι for / > 1. By induction, it can be shown that

Next, invert A_{n} = TT* _{/} , and store the result in

B_{n} : B_{n} <r- (A_{n} )^{_1} = [x • Now, set B\ - x;+ι • R,_{+}ι for i < n. Again, it can be shown that

Finally, set Ci <- B\, and - At.\ ■ Bj for i > 1. Furthermore, C\ = B\ = x ^{1}, and, by combining, C, = A,-._{\} ■ B_{t} = xf^{l} for i > 1. This embodiment has thus inverted each x,-.

Each phase above requires n - 1 multiplications, since one ofthe n values is available without recourse to multiplication in each phase. Therefore, the entire algorithm computes the inverses of all the inputs in 3n - 3 multiplications and a single inversion. In another embodiment batched division can be combined with delayed division, wherein promises at the leaves ofthe batch tree are evaluated using batched division. Consequently, only a single modular inversion is required for the entire batching procedure. Note that the batch division algorithm can be easily modified to conserve memory and store only n intermediate values at any given time.

The Chinese Remainder Theorem is typically used in calculating RSA decryptions. Rather than computing m <- v^{d} (mod N), the modulo p and q is evaluated:
^{m}P <~ ^{V}P" (^{moα}» m_{g} <- * (^{mod} ?)•

Here , = mod - 1 and d_{q} = d mod # - 1. Correspondingly the CRT can calculate m from m_{p} and m_{?}. This is approximately 4 times faster than evaluating m directly. This idea extends naturally to batch decryption. In one embodiment each encrypted message v,- modulo p and q is reduced. Then, instead of using a single batch tree modulo N, two separate, parallel batch trees, modulo p and q, are used and then combined to the final answers from both using the CRT. Batching in each tree takes between a quarter and an eighth as long as in the original, unified tree since the number-theoretical primitives employed, as commonly implemented, take quadratic or cubic time in the bit-length ofthe modulus. Furthermore, the b CRT steps required to calculate each m,- mod N afterwards takes negligible time compared to the accrued savings.

Another embodiment referred to herein as Simultaneous Multiple Exponentiation provides a method for calculating a " ■ b mod m without first evaluating a" -b^{v} . It requires approximately as many multiplications as does a single exponentiation with the larger of u or v as an exponent.

For example, in the percolate-upward step, V - V ^{κ ■} V^^{L} the entire right-hand side can be computed in a single multi-exponentiation. The percolate-downward step involves the calculation ofthe quantity _{L} ^{L} ' ^{V} _{R} ^{R} _{>} which can be accelerated similarly. These small-exponentiations-and- product calculations are a larger part ofthe extra bookkeeping work required for batching. Using Simultaneous Multiple Exponentiation reduces the time required to perform them by close to 50% by combining the exponentiation process.

Yet another embodiment involves Node Reordering. Normally there are two factors that determine performance for a particular batch of keys. First, smaller encryption exponents are better. The number of multiplications required for evaluating a small exponentiation is proportional to the number of bits in the exponent. Since upward and downward percolation both use 0(b)
small exponentiations, increasing the value of e = IT e_{t} can have a drastic effect on the efficiency of batching.

Second, some exponents work well together. In particular, the number of multiplications required for a Simultaneous Multiple Exponentiation is proportional to the number of bits in the larger ofthe two exponents. If batch trees are built that have balanced exponents for multiple exponentiation (E_{L} and E_{R}, thenXx andX^, at each inner node), the multi-exponentiation phases can be streamlined.

With b = 4, optimal reordering is fairly simple. Given public exponents e_{j} < e_{2} < e_{3} < e_{4} , the arrangement e, -e_{4} -e_{2} -e_{3} minimizes the disparity between the exponents used in Simultaneous Multiple Exponentiation in both upward and downward percolation. Rearranging is harder for b > 4.

Figure 4 is an embodiment of a system 400 for improving secure communications. The system includes multiple client computers 432, 434, 436, 438 and 440 which are coupled to a server system 410 through a network, 430.

The network 430 can be any network, such as a local area network, a wide area network, or the Internet. Coupled among the server system 410 and the network 430 is a decryption server. While illustrated as a separate entity in Figure 4, the decryption server can be located independent ofthe server system or in the environment or among any number of server sites 412, 414 and 416. The client computers each include one or more processors and one or more storage devices. Each ofthe client computers also includes a display device, and one or more input devices. All ofthe storage devices store various data and software programs. In one embodiment, the method for improving secure communications is carried out on the system 400 by software instructions executing on one or more ofthe client computers 432 - 440. The software instructions may be stored on the server system 410 any one ofthe server sites 412 - 416 or on any one ofthe client computers 432 - 440. For example, one embodiment presents a hosted application where an enterprise requires secure communications with the server. The software instructions to enable the communication are stored on the server and accessed through the network by a client computer operator ofthe enterprise. In other embodiments, the software instructions may be stored and executed on the client computer. A user ofthe client computer with the help of a user interface can enter data required for the execution ofthe software instructions. Data required for the execution ofthe software instructions can also be accessed via the network and can be stored anywhere on the network.

Building the batch RSA algorithm into real-world systems presents a number of architectural challenges. Batching, by its very nature, requires an aggregation of requests. Unfortunately, commonly-deployed protocols and programs are not designed with RSA aggregation in mind. The solution in one embodiment is to create a batching server process that provides its clients with a decryption oracle, abstracting away the details ofthe batching procedure.

With this approach modifications to the existing servers are minimized. Moreover, it is possible to simplify the architecture ofthe batch server itself by freeing it from the vagaries ofthe SSL protocol. An example ofthe resulting web server design is shown in Figure 5. Note that in batching the web server manages multiple certificates, i.e., multiple public keys, all sharing a common modulus N 510.

One embodiment for managing multiple certificates is the two-tier model. For a protocol that calls for public-key decryption, the presence of a batch-decryption server 520 induces a two-tier model. First is the batch server process that aggregates and performs RSA decryptions. Next are client processes that send decryption requests to the batch server. These client processes implement the higher-level application protocol (e.g., SSL) and interact with end-user agents (e.g., browsers). Hiding the workings ofthe decryption server from its clients means that adding support for batch RSA decryption to existing servers engenders the same changes as adding support for hardware-accelerated decryption. The only additional challenge is in assigning the different public keys to the end-users such that there are roughly equal numbers of decryption requests with each e_{t}. As the end-user response times are highly unpredictable, there is a limit to the flexibility that may be employed in the public key distribution.
If there are £ keys each with a corresponding certificate, it is possible to create a web with ck web server processes with a particular key assigned to each. This approach provides that individual server processes need not be aware ofthe existence of multiple keys. The correct value for c depends on factors such as, but not limited to, the load on the site, the rate at which the batch server can perform decryption, and the latency ofthe communication with the clients.

Another embodiment accommodates workload unpredictability. The batch server performs a set of related tasks including receiving requests for decryption, each of which is encrypted with a particular public exponent e,-.

Having received the requests it aggregates these into batches and performs the batch decryption as described herein. Finally, the server responds to the requests for decryption with the corresponding plain-text messages. The first and last of these tasks are relatively simple I/O problems and the decryption stage is discussed herein. What remains is the scheduling step.

One embodiment possesses scheduling criteria including maximum throughput, minimum turnaround time, and minimum turnaround-time variance. The first two criteria are self-evident and the third is described herein. Lower turnaround-time variance means the server's behavior is more consistent and predictable which helps prevent client timeouts. It also tends to prevent starvation of requests, which is a danger under more exotic scheduling policies.

Under these constraints a batch server's scheduling can implement a queue where older requests are handled first. At each step the server seeks the batch that allows it to service the oldest outstanding requests. It is impossible to compute a batch that includes more than one request encrypted with any particular public exponent e,. This immediately leads to the central realization about batch scheduling that it makes no sense, in a batch, to service a request that is not the oldest for a particular e,. However, substituting the oldest request for a key into the batch improves the overall turnaround-time variance and makes the batch server better approximate a perfect queue.

Therefore, in choosing a batch, this embodiment needs only consider the oldest pending request for each e_{t}. To facilitate this, the batch server keeps k
queues Q or one for each key. When a request arrives, it is placed onto the queue that corresponds to the key with which it was encrypted. This process takes O(l) time. In choosing a batch, the server examines only the heads of each ofthe queues. Suppose that there are k keys, with public exponents ei, . . ., e_{t}, and that the server decrypts requests in batches of b messages each. The correct requests to batch are the b oldest requests from amongst the k queue heads. If the request queues Q_{\} are kept in a heap with priority determined by the age ofthe request at the queue head, then batch selection can be accomplished by extracting the maximum, oldest-head, queue from the heap, de-queue the request at its head, and repeat the process to obtain b requests to batch. After the batch has been selected, the b queues from which requests were taken may be replaced in the heap. The entire process takes O(blg£) time.

Another embodiment utilizes multi-batch scheduling. While the process described above picks only a single batch, it is possible, in some cases, to choose several batches at once. For example, with b = 2, k = 3, and requests for the keys 3-3-5-7 in the queues, the one-step lookahead may choose to do a 5-7 batch first, after which only the unbatchable 3-3 remain. A smarter server could choose to do 3-5 and 3-7 instead. The algorithms for doing lookahead are more complicated than the single-batch algorithms. Additionally, since they take into account factors other than request age, they can worsen turnaround- time variance or lead to request starvation.

A more fundamental objection to multi-batch lookahead is that performing a batch decryption takes a significant amount of time. Accordingly, if the batch server is under load, additional requests will arrive by the time the first chosen batch has been completed. These can make a better batch available than was without the new requests.

But servers are not always under maximal load. Server design must take different load conditions into account. One embodiment reduces latency in a medium-load environment by using k public keys on the web server and allowing batching of any subset of b of them, for some b < k. To accomplish
this the batches must be pre-constructed and the constants associated with(_{j} ) batch trees must be keep in memory one for each set of e's.

However, it is no longer necessary to wait for exactly one request with each e before a batch is possible. For k keys batched b at a time, the expected number of requests required to give a batch is

E[# requests] = k - _{:} — .

;=_{1} k — t + 1

This equation assumes each incoming request uses one of the k keys randomly and independently. With b = 4, moving from k = 4 to k = 6 drops the expected length ofthe request queue at which a batch is available by more than

The particular relationship of b and k can be tuned for a particular server. The batch-selection algorithm described herein is time-performance logarithmic in k, so the limiting factor on k is the size ofthe &* prime, since particularly large values of e degrade the performance of batching. In low-load situations, requests trickle in slowly, and waiting for a batch to be available can introduce unacceptable latency. A batch server should have some way of falling back on unbatched RSA decryption, and, conversely, if a batch is available and batching is a better use of processor time than unbatched RSA, the servers should be able to exploit these advantages. So, by the considerations given above, the batch server should perform only a single unbatched decryption, then look for new batching opportunities.

Scheduling the unbatched decryptions introduces some complications. Previous techniques in the prior art provide algorithms that when requests arrive, a batch is accomplished if possible, otherwise a single unbatched decryption is done. This type of protocol leads to undesirable real- world behavior. The batch server tends to exhaust its queue quickly. Furthermore it responds immediately to each new request and never accumulates enough requests to batch.

One embodiment chooses a different approach that does not exhibit the performance degradation associated with the prior art. The server waits for new requests to arrive, with a timeout. When new requests arrive, it adds them to its queues. If a batch is available, it evaluates it. The server falls back on unbatched RSA decryptions only when the request-wait times out. This approach increases the server's turnaround-time under light load, but scales gracefully in heavy use. The timeout value is tunable. Since modular exponentiation is asymptotically more expensive than the other operations involved in batching, the gain from batching approaches a factor-of-b improvement only when the key size is improbably large. With 1024-bit RSA keys the overhead is relatively high and a naive implementation is slower than unbatched RSA. The improvements described herein lower the overhead and improve performance with small batches and standard key-sizes.

Batching provides a sizeable improvement over plain RSA with b = 8 and n = 2048. More important, even with standard 1024-bit keys, batching significantly improves performance. With b = 4, RSA decryption is accelerated by a factor of 2.6; with b = 8, by a factor of almost 3.5. These improvements can be leveraged to improve SSL handshake performance.

At small key sizes, for example n = 512, an increase in batch size beyond b = 4 provides only a modest improvement in RSA performance. Because ofthe increased latency that large batch sizes impose on SSL handshakes, especially when the web server is not under high load, large batch sizes are of limited utility for real- world deployment.

SSL handshake performance improvements using batching can be demonstrated by writing a simple web server that responds to SSL handshake requests and simple HTTP requests. The server uses the batching architecture described herein. The web server is a pre-forked server, relying on "thundering herd" behavior for scheduling. All pre-forked server processes contact an additional batching server process for all RSA decryptions as described herein.

Batching increases handshake throughput by a factor of 2.0 to 2.5, depending on the batch size. At better than 200 handshakes per second, the batching web server is competitive with hardware-accelerated SSL web servers, without the need for the expensive hardware.

Figure 6 is a flow diagram for improving secure socket layer communication through batching of an embodiment. As in a typical initial handshake between server and client in establishing a secure connection, the client uses the server's public key to encrypt a random string R and then sends the encrypted R to the server 620. The message is then cached 625 and the batching process begins by determining is there is sufficient encrypted messages coming into the server to form a batch 630. Ifthe answer to that query is no, it is determined ifthe scheduling algorithm has timed out 640. Again ifthe answer is no the message returns to be held with other cached messages until a batch has been formed or the scheduler has timed out. Ifthe scheduler has timed out 640 then the web server receives the encrypted message from the client containing R 642 The server then employs the private key ofthe public / private RSA key pair to decrypt the message and determine R 646. With R determined the client and the server use R to secure further communication 685 and establish an encrypted session 690.

Should enough encrypted messages be available to create a batch 630 the method examines the possibility of scheduling multiple batches 650. With the scheduling complete the exponents ofthe private key are balanced 655 and the e root ofthe combined messages is extracted 658 allowing a common root to be determined and utilized 660. The embodiment continues by reducing the number of inversions by conducting delayed division 662 and batched division 668. With the divisions completed, separate parallel batch trees are formed to determine the final inversions that are then combined 670. At this point simultaneous multiple exponents are applied to decrypt the messages 672 which are separated 676 and sent to the server in clear text 680. With the server and client both possessing the session key R 685 a encrypted session can be established 690.

Batching increases the efficiency and reduces the cost of decrypting the cipher-text message containing the session's common key. By combining the decryption of several messages in an optimized and time saving manner the server is capable of processing more messages thus increasing bandwidth and improving the over all effectiveness ofthe network. While the batching techniques described previously are a dramatic improvement in secure socket layer communication, other techniques can also be employed to improve the handshake protocol.

From the above description and drawings, it will be understood by those of ordinary skill in the art that the particular embodiments shown and described are for purposes of illustration only and are not intended to limit the scope of the claimed invention.

Patent Citations

Cited Patent | Filing date | Publication date | Applicant | Title |
---|---|---|---|---|

EP0946018A2 * | 25 Mar 1999 | 29 Sep 1999 | Nippon Telegraph and Telephone Corporation | Scheme for fast realization of encryption, decryption and authentication |

Non-Patent Citations

Reference | ||
---|---|---|

1 | * | TAKAGI T: "FAST RSA-TYPE CRYPTOSYSTEM MODULO PKQ" ADVANCES IN CRYPTOLOGY. CRYPTO '98. 18TH ANNUAL INTERNATIONAL CRYPTOLOGY CONFERENCE. SANTA BARBARA, AUG. 23 - 27, 1998. PROCEEDINGS, LECTURE NOTES IN COMPUTER SCIENCE;VOL. 1462, BERLIN: SPRINGER, DE, 23 August 1998 (1998-08-23), pages 318-326, XP000792177 ISBN: 3-540-64892-5 |

2 | * | W. STALLINGS: "CRYPTOGRAPHY AND NETWORK SECURITY: PRINCIPLES AND PRACTICE" 1999 , PRENTICE HALL , NEW JERSEY XP002219172 page 450, line 21 -page 456, last line |

Classifications

International Classification | G06F7/72, H04L9/30, H04L29/06 |

Cooperative Classification | H04L9/302, H04L63/166, G06F7/723, G06F7/721, H04L63/0442, H04L9/0841 |

European Classification | H04L63/16D, H04L63/04B2, G06F7/72E, H04L9/30F |

Legal Events

Date | Code | Event | Description |
---|---|---|---|

20 Dec 2001 | AK | Designated states | Kind code of ref document: A2 Designated state(s): AE AG AL AM AT AU AZ BA BB BG BR BY BZ CA CH CN CO CR CU CZ DE DK DM DZ EC EE ES FI GB GD GE GH GM HR HU ID IL IN IS JP KE KG KP KR KZ LC LK LR LS LT LU LV MA MD MG MK MN MW MX MZ NO NZ PL PT RO RU SD SE SG SI SK SL TJ TM TR TT TZ UA UG US US US UZ VN YU ZA ZW |

20 Dec 2001 | AL | Designated countries for regional patents | Kind code of ref document: A2 Designated state(s): GH GM KE LS MW MZ SD SL SZ TZ UG ZW AM AZ BY KG KZ MD RU TJ TM AT BE CH CY DE DK ES FI FR GB GR IE IT LU MC NL PT SE TR BF BJ CF CG CI CM GA GN GW ML MR NE SN TD TG |

13 Feb 2002 | 121 | Ep: the epo has been informed by wipo that ep was designated in this application | |

21 Nov 2002 | REG | Reference to national code | Ref country code: DE Ref legal event code: 8642 |

16 Jul 2003 | 122 | Ep: pct application non-entry in european phase | |

5 Apr 2005 | NENP | Non-entry into the national phase in: | Ref country code: JP |

Rotate