China
Africa
Explore chapters and articles related to this topic
Communicating with Other Xlets
Published in Steven Morris, Anthony Smith-Chaigneau, Interactive TV Standards, 2012
Steven Morris, Anthony Smith-Chaigneau
As we have already mentioned, one approach is to use a hash table to map imported objects to their stub implementations in the client application, but this means that both objects have an unnecessary reference that may prevent them from being garbage collected. To avoid this, references to these objects within the RMI implementation should always be weak references. A weak reference is a Java reference that is not counted by the garbage collector, and thus any object referred to using only weak references will be garbage collected. These are not officially available in version 1.1 of the Java platform, but some implementations of Java 1.1 may support them as an extension.
Aspect-oriented challenges in system integration with microservices, SOA and IoT
Published in Enterprise Information Systems, 2018
Dealing with ever-changing data representations in distributed systems is a challenge noticed by Cerny and Donahoo (2016). This article suggests that the knowledge of data representation should be encapsulated in the component responsible for persistence. Moreover, this component should provide another component with meta-information about the data structure. This approach accommodates existing enterprise development standards (Cerny and Donahoo 2015) that promote capturing validation or even security concerns bound to data representation through annotations. Streaming other concerns together with data structure meta-information enables integrating components to integrate structure, security, and validation on the fly and thus automatically adapt to changes. However, the integrating components only weakly reference such data components as proxies since a stronger reference would lead to coupling and dependency. An example usage of a weak reference where it is applied in practice is user interface (Cerny and Donahoo 2015). The considerable advantage of the approach (Cerny and Donahoo 2015) is that user interface can be presented on different platforms using native components, while only providing a single backend. The aspect weaver is thus distributed between backend and frontend, leading into automated adjustments in all the heterogeneous user interfaces upon a backend change. For instance, when a novel data field is introduced on a backend component, e.g. with novel validation rules and constraints, no redesign of the user interface is needed as it automatically appears in the update structural stream.