Font Size: a A A

The VLSI Implementation Of Elliptic Curve Cryptography IP

Posted on:2005-09-24Degree:MasterType:Thesis
Country:ChinaCandidate:X F TangFull Text:PDF
GTID:2168360122975046Subject:Circuits and Systems
Abstract/Summary:PDF Full Text Request
With the rapid development of computer technology and network, information security is given more and more attention. As new public-key cryptography, elliptic curve cryptography (ECC) has many excellent attributes: short key length, fast speed and high security. These attributes make it an ideal choice for application. From the point of algorithms and hardware, this article discusses the VLSI implementation of ECC IP.The mathematical background of ECC is very complex and many algorithms involved. Based on the observation of the character of irreducible polynomial in characteristic two finite fields, we proposed a fast reduction method. By this method, the speed of modular multiplication improved greatly. Following the discussion of the design preferences of ECC, we suggest that a symmetric arithmetic unit will be better. Then algorithms based on symmetric arithmetic unit are discussed. We proposed a dual field multiplier constructed from dual field adder and Wallace Tree. By this multiplier, dual field ECC is supported with small area and less energy assumption. In order to reduce the path delay and enhance the system frequency, we suggest the four-stage pipelined data path organization. Corresponding to it, a pipelined data path controller is suggested. And we discussed the two solutions for data dependency: forwarding and pipeline stalling. Instruction sets is suggested because of three reasons: reducing the complexity of the algorithms' control unit design, enhance the flexibility and expansibility. Then we discussed the design principal of instruction sets, instruction format and instruction FIFO. The call between algorithms' control blocks as well as communication between algorithms' control block and instruction FIFO is also discussed. In order to avoid the bottleneck caused by RAM, two RAM each with two asynchronous read ports and one synchronous write port are used. The system we implemented is very flexible and universal. Furthermore, high performance is gained with small area.
Keywords/Search Tags:Elliptic Curve Cryptography, Point multiplication, Montgomery method for modular multiplication, Fast reduction method, Dual field multiplier, Pipeline
PDF Full Text Request
Related items