When planning to deploy an Oracle database that uses ScaleIO, the solution architect or DBA must decide if the ScaleIO cluster will be converged or hyperconverged.
With a converged architecture, as shown in Figure 1, each node in the cluster will either be a ScaleIO Data Client (SDC) consuming storage from ScaleIO and running the database software or a ScaleIO Data Server (SDS) storing ASM data and providing it to the SDC/database nodes.
In a hyperconverged architecture, as shown in Figure 2, some or all nodes will be both SDCs and SDSs, both running the database software and acting as storage nodes for ScaleIO.
The hyperconverged architecture may allow for more efficient utilization of the compute, network and storage resources, but may require additional Oracle software licenses.
Keeping SDCs and SDSs separate may lead to some nodes being less than fully utilized, but may be beneficial from a license management perspective, while providing for more I/O headroom.