⭐⭐⭐⭐⭐ Performance Comparison of Database Server based on #SoC #FPGA and #ARM Processor
Download as PPTX, PDF0 likes1,087 views
The document presents a performance comparison of database servers based on System on Chip FPGA and ARM processors, highlighting a proposed solution to run a MySQL database server on embedded Linux using FPGAs. Benchmark tests indicate that FPGAs significantly reduce response times for simultaneous client requests compared to a Raspberry Pi, which experiences longer response delays. Future work involves the implementation of a gateway for edge-fog computing to enhance response times in sensor networks.
![Related Work
• Embedded Linux can run on FPGAs together with several IoT applications, such as a database
server, web server, DNS server, traffic analyzer, among others. A lot of related work has been
done evaluation of query performance [6], delay minimization [7,8] and features based on
speeds and operating time [9,10].
• Lee et Al. [6] performed benchmark tests with SQLite to evaluate the use of FPGAs together
with DRAM/PRAM hybrid memories (SmartSSD) in order to offload the processing to the SSD.
The authors demonstrated their proposal outperforms the CPU-based approach.
• In [7], the authors proposed a configuration with interaction between the HPS, FPGA with
peripherals such as LEDs or switches DE1-SoC FPGA and an ARM Cortex-A9 processor. FPGA
has applications in systems where considerable amounts of data are processed with low
latency.
• Wielgosz and Karwatowski described the importance of having an optimal latency level in a
database system [8].](https://image.slidesharecdn.com/ieee-latinamericanconference-comparisonofdatabaseserverbasedonsocfpgaandarmv2-211227040337/95/Performance-Comparison-of-Database-Server-based-on-SoC-FPGA-and-ARM-Processor-7-638.jpg)
![Discussion and Conclusions
• In this paper, we proposed a solution using FPGAs to run a MySQL database server on embedded
Linux due to the fact that this device can be used in real-world applications that involve sensors to
measure environmental parameters.
• Available benchmarking tools were used to benchmark the service running on two different
development boards, namely FPGA and Raspberry PI 4B +. It was found that using an FPGA as a
database server allows us to reduce the response time of multiple clients that make simultaneous
requests to the system thanks to its hardware capacity without excessive CPU and memory usage,
while the Raspberry PI requires between a25 % and 50 % longer than FPGA’s response time.
• As future work, we propose to implement a gateway to perform Edge-Fog computing based on a
Raspberry-Pi computing module in order to improve the response time of sensor networks to the cloud.
In fact, the proposed architecture can be applied to add the edge database server and to implement
fast and intelligent control algorithms with sensor networks for precision agriculture [12] or turkey
farming [13].](https://image.slidesharecdn.com/ieee-latinamericanconference-comparisonofdatabaseserverbasedonsocfpgaandarmv2-211227040337/95/Performance-Comparison-of-Database-Server-based-on-SoC-FPGA-and-ARM-Processor-12-638.jpg)
⭐⭐⭐⭐⭐ Performance Comparison of Database Server based on #SoC #FPGA and #ARM Processor
- 1. Performance Comparison of Database Server based on SoC FPGA and ARM Processor Escuela Superior Politécnica del Litoral, ESPOL, Guayaquil, Ecuador Centro de Tecnologías de Información, CTI Facultad de Ingeniería en Electricidad y Computación, FIEC Rebeca Estrada Pico , Víctor Asanza , Jocelyn Miranda , Leiber Rivas , Danny Torres
- 3. When using this resource, please cite the original publication: V. Asanza, R. Estrada, J. Miranda, L. Rivas and D. Torres, "Performance Comparison of Database Server based on SoC FPGA and ARM Processor," 2021 IEEE Latin-American Conference on Communications (LATINCOM), 2021, pp. 1-6, doi: 10.1109/LATINCOM53176.2021.9647742. Source code repository: https://github.com/jocammir/Sistema_gestion_base_de_datos_FPGA_HPS_DE10Standard
- 4. Topics • Introduction • Related Work • Dataset • Methodology • Results • Discussion and conclusion Performance Comparison of Database Server based on SoC FPGA and ARM Processor
- 5. En el 2050 … Introduction
- 6. Introduction
- 7. Related Work • Embedded Linux can run on FPGAs together with several IoT applications, such as a database server, web server, DNS server, traffic analyzer, among others. A lot of related work has been done evaluation of query performance [6], delay minimization [7,8] and features based on speeds and operating time [9,10]. • Lee et Al. [6] performed benchmark tests with SQLite to evaluate the use of FPGAs together with DRAM/PRAM hybrid memories (SmartSSD) in order to offload the processing to the SSD. The authors demonstrated their proposal outperforms the CPU-based approach. • In [7], the authors proposed a configuration with interaction between the HPS, FPGA with peripherals such as LEDs or switches DE1-SoC FPGA and an ARM Cortex-A9 processor. FPGA has applications in systems where considerable amounts of data are processed with low latency. • Wielgosz and Karwatowski described the importance of having an optimal latency level in a database system [8].
- 8. Dataset
- 9. Methodology
- 10. Methodology
- 11. Results
- 12. Discussion and Conclusions • In this paper, we proposed a solution using FPGAs to run a MySQL database server on embedded Linux due to the fact that this device can be used in real-world applications that involve sensors to measure environmental parameters. • Available benchmarking tools were used to benchmark the service running on two different development boards, namely FPGA and Raspberry PI 4B +. It was found that using an FPGA as a database server allows us to reduce the response time of multiple clients that make simultaneous requests to the system thanks to its hardware capacity without excessive CPU and memory usage, while the Raspberry PI requires between a25 % and 50 % longer than FPGA’s response time. • As future work, we propose to implement a gateway to perform Edge-Fog computing based on a Raspberry-Pi computing module in order to improve the response time of sensor networks to the cloud. In fact, the proposed architecture can be applied to add the edge database server and to implement fast and intelligent control algorithms with sensor networks for precision agriculture [12] or turkey farming [13].
- 14. For more information Mail: {restrada, vasanza, jocammir, lvrivas, daaltorr}@espol.edu.ec Facultad de Ingeniería en Electricidad y Computación, FIEC Escuela Superior Politécnica del Litoral, ESPOL Campus Gustavo Galindo Km 30.5 Vía Perimetral, P.O. Box 09-01-5863 090150 Guayaquil, Ecuador Rebeca Estrada Pico , Víctor Asanza , Jocelyn Miranda , Leiber Rivas , Danny Torres
- 15. Thank you!