Page 51 - Kaleidoscope Academic Conference Proceedings 2024

Innovation and Digital Transformation for a Sustainable World

3.5 Upgrade cycle evaluation

From the perspective of the upgrade cycle assessment,
assuming customers add n new inspection rules every day,
the average upgrade iteration cycle in the production
environment is 30 days. Suppose the upgrade iteration is
represented by formula (14), where “a” is the amount of code
or configuration to be modified for inspection rules (one line
of code is equivalent to one line of configuration), “b” is the
amount of compilation work required for the change, and “c”
is the complexity of the deployment.

= ( ) + ( ) + ( ) (14)
Figure 7 – Resource pool data with significant
improvement in data quality
The implementation approach of the engine incurs a
complexity of 1 for each new rule, so the complexity for n
3.3 Data incremental efficiency
rules is n. Since it is implemented through configuration,
there is no need to modify the code or recompile, resulting in
In testing environment, we repeatedly executed inspection
tasks for different volumes of host data using 5 rules. We a workload of 0 for this part. Without the need to modify the
code, and implementing through configuration, it is possible
measured the execution time of the inspection code based on
expression engine at different data volumes, as shown in to upgrade to the production environment in real-time,
provided that the inspection expressions are ready. Therefore,
Figure 8(A). It’s not difficult to observe that the engine
execution time is almost linearly related to the data volume using the engine approach, the overall iteration complexity
is simplified to formula (15):
and doesn’t exhibit exponential growth. The experimentally
obtained processing times are also fully capable of meeting
the requirements of the production environment. = (15)
The approach of implementing directly through logic, for
each new rule, requires an average addition of 5 lines of code,
resulting in a complexity of 5n for n rules. As code
modification is necessary, b(n) is not equal to 0. Since
modification of the program is needed to implement new
functionalities, and the iteration cycle is 30 days, the
complexity of c(n) is between 1 and 30. Therefore, using the
approach of direct logic implementation, the overall iteration
complexity is expressed in formula (16):
Figure 8 – (A) Execution time of inspection code based
on expression engine under different data volumes. (B) = 5 + ( ) + , ∈ [1,30] (16)
Comparison of expression engine code volume and
logical code volume 4. CONCLUSION

3.4 Code efficiency comparison This paper proposes a method based on the native
functionality of an expression engine and business
After multiple practical tests, such as testing various rules in extensions in different scenarios. This method enables
the CMDB test environment, including 20 rules for bare dynamic inspection of business data, effectively addressing
metal, 13 rules for host machines, and 7 rules for virtusl challenges such as diverse data sources, complex data
servers, the experimental results indicate that the basic code analysis, and data management. Using the CMDB business
volume of the expression engine is approximately 180 lines. scenario as a case study, we apply the expression engine-
For each additional rule, the expression code increases by 1 based inspection method to the production environment,
line (we maintain the expressions in Mysql for ease of ultimately implementing a dynamic inspection solution for
repeated configuration). In contrast, using direct logic code various models in CMDB. Through the analysis and tracking
for inspection has a basic code volume of about 50 lines, and of data reported by 49 product sources and optimization by
for each additional rule, an average of 5 lines of code needs the reporting party, the reporting accuracy stabilizes at over
to be added. As the number of inspection rules increases, the 90%, with an average improvement of approximately 12%.
overall code volume trend is shown in Figure 8(B). When the This inspection method can meet customer business
number of rules is 100, the code volume of this method is requirements without code modification and frequent
reduced to 50% of the logical code. As the number of rules production environment cutover, particularly in scenarios
continues to increase, the code volume of this method with evolving business needs. Compared to traditional logic-
eventually approaches 20% of the logical code, and the code based inspection methods, the expression engine-based
reduction amount ultimately approaches 80%. inspection method significantly reduces iteration complexity.

– 7 –