"The role of humans as database administrators will cease to exist. These future systems will be too complex for a human to reason about. DBMSs will finally be completely autonomous and self-healing"
Dunno, seems like as long as there is crappy data that humans need to clean, enterprise and financial firms will continue to use XL as critical part of their data infrastructure.
And as long as XL reigns supreme in finance and consulting, seems a bit far fetched to talk about infinitely scalable, sentient and 'self-healing' DBMSs...
Let's focus on getting the data out of XL, then work on the genie in the bottle.
Pavlo's research on "autonomous databases" is focused less on issues of data quality, and more around configuration and tuning of database systems [1]. For most databases, there are a staggering number of knobs to tune and configure for a particular workload. DBAs typically work on this type of tuning, whereas data quality issues are often problems at the application layer or above.
In my personal experience most of the tuning of a database happens in tandem with tuning the application. E.g. you find a set of problematic queries and then go back to the application to see if they should be fixed or if the database is incorrectly tuned. Quite often it is a mix of both, and just automating most of the tuning only just changes the role of the DBA, not does away with it.
Dunno, seems like as long as there is crappy data that humans need to clean, enterprise and financial firms will continue to use XL as critical part of their data infrastructure.
And as long as XL reigns supreme in finance and consulting, seems a bit far fetched to talk about infinitely scalable, sentient and 'self-healing' DBMSs...
Let's focus on getting the data out of XL, then work on the genie in the bottle.