Fig. 7. Response times for scaling parallel requests.
Furthermore, with the exception of POST (xml to json-ld) in the trials,
the CIM is capable of converting payloads without significant
statistical difference in the majority of situations. This exception
happens when the payload is translated by the first CIM; while the
remote CIM is translating, this exception does not occur; as a result of
this discovery, multithreading in the first CIM is most likely impeding
data translation.
Experiment 3: scaling a payload size
The GET (json-ld) and POST (json-ld) flows are shown in Figs. 5 and 6,
respectively. In addition, a 1 Mb payload was synthetically enlarged to
9 Mb by 0.25 Mb pieces. Fig. 8 shows the times it took the CIMs to
complete the GET and POST requests, depending on the size of the
payload. The time it took to complete both queries in seconds is the
performance statistic in this experiment. Finally, for security
considerations, the CIM limits the size of the payload that may be
exchanged to 9Mb; as a result, the maximum size in the experiment is the
aforementioned.
Based on the data shown in Fig. 8, it can be determined that the CIM
takes less than a half-minute to handle requests involving 9 Mb
utilizing GET or POST requests; nevertheless, payloads involving 1 Mb
require around 1 s to solve.
Conclusion: The CIM can handle payloads up to 9 Mb in a fair amount of
time, as seen in Fig. 8. Note that there is a security constraint in
place to prevent DoS attacks and server congestion caused by message
overflow. Consider that most DR systems do not require huge volumes of
data to be sent. at least not rapidly; instead, they must generally
exchange tiny messages relatively quickly. As a result, even if the CIM
takes seconds to respond to huge payloads, it can be said that it meets
the data interchange criteria for DR systems in terms of size [49].