| Popis: |
Research clouds contain diverse and large datasets, this data is published by the use of Persistent Identiers (PIDs). The current paradigm utilizes data transmissions by the means of host-to-host connections (IP), where every data request from the consumer is answered with a data transfer from the source (the producer). This approach can potentially cause congestion and delays with many data consumers. Named Data Networking (NDN) is a data centric approach where unique data, once requested, is stored on intermediate hops in the network. Consecutive requests for that unique data object are then made available by these intermediate hops (caching). This approach distributes trac load more ecient and reliable compared to host-to-host connection oriented techniques [69]. Furthermore, one of the most important technical challenges is to incorporate interoperability between NDN and the dierent PID naming schemas. These naming schemas are used by data providers within these data infrastructures for sharing and identifying digital objects. This research investigated how identication services and transmission services can be better integrated by the use of NDN. In order to create this integration, a translation between the dierent naming schemas was developed and made extendable for future PID types. Furthermore, a method was researched and applied for planning and scaling an NDN with scalability in mind. |
