EMC's new all-flash array, named XTremIO, will be launched next month. In terms of the market segment for this array, Oracle would certainly be a big target. See, for example, this white paper by the Uptime Institute on XTremIO performance in the context of Oracle and SAP.

An all-flash array is a killer feature for Oracle, especially in the case of a workload that looks like this:

• Large amounts of small block, random I/O
• Read intensive workload
• I/O latency is a performance bottleneck
• I/O bandwidth is a performance bottleneck

• The database will fit into the capacity of an all-flash array (each XTremIO brick has a capacity of 7 TB usable)

It is impressive how many Oracle database workloads look like that. Pretty much any heavily-used production OLTP database (unless it is breathtakingly large) would benefit from XTremIO. Assuming sufficient I/O activity (and what production Oracle database have you seen recently that is not I/O stressed?), an all-flash array will definitely help performance.

Given the costs of running Oracle, any significant performance improvement will return a very nice ROI. An all-flash array is certainly no exception. For this reason, we fully expect for the XTremIO all-flash array to be a huge hit in the Oracle space.

Do you plan to implement XtremIO all-flash arrays for your Oracle production workloads? We anticipate that training for XtremeIO will be available early next year.