What Is Your Checklist to Identify Owners of "Shared Resources"

[Robert Campbell]

We keep revisiting the conversation of who owns "shared" data. We are tagging everything we can but when it comes to things like databases someone always refers to it as a shared database. One approach was to "assign" it to the application that the database gets deployed with even though other applications will be using it. However, we've been having may releases that involve multiple applications that use the same database so that approach no longer works. And because the data that's inserted into the database comes from multiple apps, we can't assign it to the app that writes to it.  There were some other approaches presented but I'm curious how others are managing this.

Are you aligning it with one application or are you tagging it with multiple applications? There are other resources such as elasticache and S3, etc.

[Niklas Peterson]

Hi,

I think the trick is to break it down to a level where the "shared data" is a service itself. A service with the purpose of providing the "shared data" to other services.

With that approach you get the typical provider-consumer relationship and a defined ownership.

Regards,
Niklas

[Robert Campbell]

We do this for our "data warehouses" but that is easy because there is a single team that owns that. My suggestion is to assign it to the org/dept/domain that covers the cost but the problem with that now is that these org/dept/domains have multiple apps that use the same db. So even if we did provide it as a shared data service, we would need to identify which "group" owns it. When it's under one org/dept/domain that's fine but when it's across many it's not so easy. For those shared resources, one area is billed for it and they bill other areas for their use of it so for me, the area that gets that initial bill is my suggestion but for some reason they don't like that idea.