The first one makes the point of Hana is in-memory vs Sybase IQ being on disk. They also mention as of Hana 1.0 there is no DR/High Availability options. When you have this massive in-memory database think about how do I transfer hundreds of GB of memory to another machine and keep it synchronized -- that is a very real problem for in-memory databases. Also they allude to the fact that Hana would not scale beyond a certain point in time.