data hub data’distribution: Network Centric architecture or Hub and Spoke architecture
data’distribution comes in two architectures: Network Centric or Hub and Spoke.
The advantages of the centralized model using a data hub and a central data repository come into their own in terms of data’distribution administration, source system performance and service continuity. data’distribution can also be used to perform replications in non-connected mode.
Centralized Hub and Spoke architecture
The Hub and Spoke architecture is a centralized model with one or more central systems dedicated to data'distribution through which all communications are conveyed (data hub). The role of the hub is to centralize all the information.
The model also includes a central repository that groups together all the necessary data. All data exchanged must go through the data hub; without it, there can be no flow. A spoke is a branch (connector) linked to the hub. This model corresponds to the centralized "Publish and subscribe" exchange method, one in which the data first goes through the data hub before being transferred to the recipient.
The main benefit of a centralized system lies:
Network Centric architecture
Conversely, the Network Centric architecture is based on a decentralized model with direct communications between sources and targets. It is built on sources, each containing their own repository, relayed to targets (nodes) which are the points of connection to an application.
The advantage of this architecture is that it distributes all the data to all the sources.
Synchronizing all the repositories on a single repository
data’distribution can thus be used to synchronize all the repositories and consolidate them into a single repository, thus making up for its difference with the Hub and Spoke model.
With data’distribution, you can choose the architecture type best suited to your various criteria.
Given that all source databases are potential targets of other sources, and that all targets are potential sources for other targets, the replication is extremely flexible and can be carried out in any direction and even bi-directionally with no restrictions.
All communications are carried out via the TCP/IP protocol using a single, fully configurable IP port and no standard port. Fully firewall-controlled flows from a production network can thus be securely replicated to a demilitarized zone (DMZ) without any opening.
Replication in non-connected mode
Replication in non-connected mode by data’distribution resolves the issues faced by companies with multiple remote sites that are not inter-connected via a general network and/or with a mobile workforce. It is simply not possible to establish an unlimited number of communications.
Besides, for security reasons, it may be necessary to avoid direct interconnections between the production source on the one hand, and remote targets on the other. Replication in non-connected mode, which is similar to log shipping mode, uses a source to load a binary file containing the I/Os for the targets. When this file is physically transmitted to the targets, it can be applied locally.
data’distribution's physical architecture comprises 3 parts:
Security profiles are used to restrict the actions of developers and users who define the parameters.