http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
Performance Improvement of Hybrid Search Engine for Wide and Sparse Rules
Keiko Endo,Yuji Yano,Shingo Ata,Ikuo Oka 대한전자공학회 2017 대한전자공학회 학술대회 Vol.2017 No.1
Network equipment uses hardware-based search methods for high-speed classification. In future networks, it is expected search hardware needs to find an entry from a large table which contains complex rules. Today, TCAM (Ternary Content Addressable Memory) is used widely as a search hardware in high-speed routers. However, it is difficult to increase the capacity of TCAM. We have proposed MDSE (Multi-Dimensional Search Engine) architecture that can search the large table with power saving. MDSE can be configured by a combination of the different search components. A component of MDSE needs search binary values in power saving. In this paper, we propose TACT (The Advanced Conceptional TCAM) structure optimized for binary long key search. TACT can be applied to a search block of MDSE. We implement MDSE configured by TACT on FPGA, and evaluate that the power consumption of TACT is 7.63 mA per 4 Kbits. Furthermore, we propose TT-HSE (TCAM + TACT Hybrid Search Engine) as MDSE configured by TCAM and TACT. We show that TT-HSE reduces the power consumption by 57.9% compared with MD-TCAM, using TACT performance and the table data generated by Classbenchv6 dataset.