This implementation is highly influenced by the technology that is being used, because each one has a different way to provide access control. For example, Apache Spark uses Kerberos to perform the authentication. Actually, these access control rules are the implementation of security solutions that were previously defined in the SO component (with a higher level of abstraction); this is why in this phase there is not a specific activity for that purpose. Indeed, this implementation can be also helped by using security patterns. In this phase, it is important to identify the different users and roles that will interact with the Big Data ecosystem, as this can also influence the definition and implementation of access control rules.