A major part of the Oracle strategy around big data is focussing on how to funnel big-data into subsets which can be loaded into a data warehouse so it can be analysed. As can be seen from the below example everything is to be funnelled into a data warehouse in this approach and then OBIEE or in-database analytics can be used to analyse this subset of the original acquired data.
Within this strategy a couple of options to implement this are available. For once you can use a full Oracle engineered systems approach as shown in the image below or you could go for a solution in which you design and create parts yourself. In the below image you see how the Oracle Big Data Appliance is used to acquire and organize the data that is collected from a number of sources. From the Oracle Big Data Appliance the pre-processed data is stored within databases inside an Oracle Exadata engineered system. Finally analysts will connect to the Oracle Exalytics machine to make use of their standard Oracle BI tools to query the subsets stored within the Oracle Exadata.
One of the advantages of the above outlined solution is that the systems are designed to work together and the time to create a working solution from a infrastructure perspective is short. Also the systems are designed to provide you with a optimised performance. Also the communication between the components is handled by infiniband, for more information on how to connect exadata and exalytics together with infiniband please do refer to my presentation on this subject on slideshare.