Autonomous apps: who’s in control?
With the release of Oracle’s 18c database, which the maker has said is the first ever autonomous database, a question is raised as to control.
When an application is self-tuning, self-regulating and self-determining in terms of patching and security, who is really in control of it?
Is the maker in control, or is it the application owner or administrator?
To what extent are these self-governing features exposed to control by the latter to allow them to bring applications into line with organisation policy?
“It is probably a good thing that this debate opens with an enterprise level application such as the database, and especially from a vendor that has the pedigree of being the very foundation of many a global enterprise, but still a debate it is likely to be”
If these options are actually exposed to control by administrators, are they negating the benefits in the first place, of having an application autonomous? If an administrator can switch of self-protection, why have it in the first place?
What if a patch is autonomously applied that is critical for, say a database, but results in some incompatibility with some other layer in an software stack?
Is this a first for blogs to be written entirely in questions?
However, the point is valid in that if something is looking after itself, then by definition, an administrator is not in total control of it. There is no suggestion that the makers of what are soon to be an array of applications that are autonomous are being in any way sinister about it, but by letting something look after itself, a degree of control is lost. And, in fairness, the programming to allow something to look after itself is not going to be accomplished by those who purchase it.
So this presents something of a quandary, I’m sure.
A quandary for the vendors to communicate just how autonomous applications can integrate with existing ‘dumb’ applications and administrators who need to convince sceptical colleagues that their self-governing applications are not going to do something in their own interest that will result in hobbling something else.
It is probably a good thing that this debate opens with an enterprise level application such as the database, and especially from a vendor that has the pedigree of being the very foundation of many a global enterprise, but still a debate it is likely to be.
As we have seen the issue of control fade from the top lists of inhibitors for IT decision makers when it comes to cloud computing, we may see in similar fashion surveys and reports in the near future that list some revolving top three inhibitors for autonomous application adoption, with control being a perennial. Or we may not.
I look forward to the next milestone in intelligent, self-governing, self-determining applications when they become, I don’t know… maybe ineffable.
Now that might present a marketing challenge.