You probably already know that there is a feature for XML data processing in DB2 which is called pureXML. You also know there is a continuous availability and scale-out feature for DB2 called pureScale. And I won't talk about pureQuery today. But did you know that pureXML is supported on pureScale? Did you know that you can request access to the so-called pureScale Acceleration Kits and try it out yourself (for free)?
I recently was instructor for a pureScale Workshop and one of the attending companies reported that they built their own small pureScale system. Such a mini system - we often refer to them as nanoClusters (or here) - has all the features of a real cluster, but only costs few hundred dollars/Euros in hardware and you have your own pureScale system. If you want to "go high-end", you could request a proof of concept/proof of technology at one of the IBM locations. But in either case, test drive pureXML on pureScale...