Performance analysis of ternary content addressable memory (TCAM)

Abstract

With the advancement of scientific achievement, electronics sector has proved a radical improvement both in terms of efficiency and high speed applications. TCAMs are widely preferred in network routers for packet forwarding and packet classification. However it is effective for other applications that are suitable for high-speed table which is used for lookup. Reduction of power consumption is one of the biggest-challenges in TCAM design when it is expected not to decrease the speed or memory density. During match detection, switching of highly capacitive Match lines (ML) and Search lines (SL) cause huge power consumption. Energy efficient match detection or ML sensing schemes tries to reduce this power consumption. In this paper, different ML sensing schemes are compared in order to find out the better scheme in terms of better power and speed performance. In conventional scheme large amount of energy is wasted as number of mismatched word is higher. Several further modified schemes are discussed and compared later on. The dedicated comparison circuitry of proposed ML sensing schemes reduces power consumption by minimizing search time and limiting voltage swing of MLs. A 16×16 bit TCAM powered by 1.8V supply voltage is designed with 0.18μm CMOS for this purpose.