CONCLUSION
The most important factors affecting performance in learning-based
systems are the number of data and data diversity. In running
learning-based systems at distributed points in a system, accuracy
performance may not reach the desired level due to data set diversity
and insufficient amount of data. Collecting, storing and training data
in a central location brings certain risks and precautions.
When the features of the infrastructure proposed in this study are
compared with the studies carried out in the literature, Table 6 shows
the advantages of our infrastructure. GDPR policies, log immutability,
storage strategy, edge computing system prepared for network performance
stand out as important positive differences. In addition, basic features
such as data security, non-tamperability, distributed structure, and
identification, which are also present in other studies, are also
available in our infrastructure.
Table 6. Comparison of the use of Blockchain to secure data in the cloud
ecosystem.