Abstract:
Fast, reliable, and secure data processing is
critical in the Information Technology (IT) sector as it is an
indispensable requirement of all the growing technologies.
Achieving this requires handling large amounts of data and
massive computing power. Emerging technologies use devices
under low-performance environments, such as raspberry-pi,
which require highly efficient databases under low computing
power. Inappropriate selection of databases leads to
considerable resource wastage in numerous ways, such as waste
of time, cost, and energy. The selection of a reliable, userfriendly database compatible with low-performance
environments will downplay the energy requirement and cut
down applications' carbon footprint. Thus, this study was
conducted to compare the performance and query cost of using
ClickHouse and MySQL databases as column-oriented and roworiented databases in a low-performance environment. The
effectiveness of using column-oriented or row-oriented
databases with large amounts of data at different platforms was
analyzed. The study was conducted by running different queries
using an automated PHP script and checking the execution time,
processing power, disk I/O, and memory usage for each query.
The study was conducted with a virtual Linux environment of
1024MB RAM, 1.8Ghz single processor, and 25Gb storage,
representing very low computing power, even lesser than the
recommended minimum requirement of ClickHouse and
MySQL databases. Both databases performed well in the above
mentioned environment. Our experimental results show that
ClickHouse execution time and disk write speed is tremendously
low compared to MySQL. To obtain this performance,
ClickHouse used more CPU processing power and memory
utilization than MySQL.
