The main problem with databases solutions today is the ability to scale capacity. We don't want to be able to store terabytes of data on a single database/node, because it won't help performance. In my opinion a good innovation would be the ability to hide sharding from the developer, make the application scale in capacity by adding nodes. It's not an OPTION, it's a MUST HAVE solution for future databases. There are plenty of HA solutions out there, but NO ONE tackles the capacity proglems because now one want to mass up with sharding.
Furthermore, it's not enough, we need backup and restore options to be able to store easily backup and restore data. So many data base solution, all are far away from solving those problems. I've heard and read tons of arguments, all crap. It seems people just want to make money on their current solutions and don't bother to innovate. In today's mobile applications and social networks it's necessary to have a DB solution that can easily scale in capacity to terabytes of data using sharding behind the scene, allows us to query the database as if it was a single database and offer easy back and restore options, no in the logical save to SQL/Gzip way, but another way that will make it quicker, like detaching replicas and use binary copy to save the data.
I've spent months reading tons of information. I just can't believe that Amazon nor Microsoft do anything about it. This solution might come as open source from developers who cares about developers. It seems that this market is doomes when it comes to innovation. The only solutions that I find appropriate is ScaleBase and DBShards, which allows you to start with a small MySQL project, but when you need to scale, you pay for their services and you don't need to make any changes to your project. It seems MySQL is the better choice because companies work to find solutions for that popular DB platform, so developers can start with Amazon RDS and if and when they grow big they know that their are solutions out there. No all apps become succesfull so it's not smart to worry so much ahead of time. And if they do become successful, their are pretty good solution out there.