Enhancing Regular Expression Processing through Field-Programmable Gate Array-Based Multi-Character Non-Deterministic Finite Automata

Round 1

Reviewer 1 Report

The article shows the great benefits of using FPGAs over CPU for regular expression search.

The main drawback of this article is the lack of comparison with other implementation results (papers) in both the CPU and FPGA. The references were only used at the beginning of the paper.

I don't fully understand the comparison operation in Fig. 10. For example, how to check whether a character is a number or a lowercase letter. This comparison requires the use of LUT memory?

It is not known how optimized the code was for the CPU. It is not fully explained why the implementation of 4-NFA takes less area than 4x 1-NFA: is it a shared input-output interface e.g. PCI-Express?

How fast the regular expressions (tokens) can be changed: both CPU time (pattern to NFA in Fig. 2) and FPGA time.

Author Response

Please refer to the Word file below for details.

Author Response File: Author Response.pdf

Reviewer 2 Report

The explanation of proposed FPGA approach is more required. 

Include quantitative results to be included  in conclusion and also compared with existing techniques 

Check the  Grammatical errors 

Author Response

Please refer to the Word file below for details.

Author Response File: Author Response.pdf

Reviewer 3 Report

This research shows the efficiency and accuracy of FPGA-based processing for regular expressions compared to traditional software methods, particularly in real-time data processing. It suggests FPGA technology as a compelling alternative for complex data tasks without significant resource consumption increases, advocating for its broader adoption and suggesting future directions for exploring more advanced pattern matching algorithms and integration into database management systems.

Overall, this paper was written in English well. However, figures must be revised in terms of readability. Please double check to use a larger font for all figures. For example, in Figure 7, legends including Token Table and some labels such as Token Matcher_1 can be more readable when you use a larger font.

Author Response

Please refer to the Word file below for details.

Author Response File: Author Response.pdf