[louis-paul]

Recently discussed with somebody at a major cloud provider and IIRC their response was in essence “I think we’re doing them a favor: ask Neo4j about our release of [hosted graph database service]. We’ve taken a share of their business, but the overall market is much bigger, and they’re doing great. Our hosted products are not even the most polished in the market, and companies like Elastic can easily be compete with us.”

Thoughts?

[gregwebs]

Yes, some cloud providers readily suggest competing with them by making something better. In the end, the DB is still running on their infrastructure (required to keep the customer's latency low) and they are still capturing profit even if it is not as much as their DBaaS profit.

There is a problem though that a DBaaS provider is still at an inherent disadvantage. You aren't in the AWS/GCP cli tool and APIs. And the DBaaS has to setup VPC peering across accounts, which has some inherit limitations (e.g. no DNS resolution on GCP).

[merb]

> (e.g. no DNS resolution on GCP)

actually you do not need DNS resolution on GCP since you can actually have a high available IP way easier (which is way better). in AWS it's also possible to have a highly available IP, but it's way harder to do so.

[bch]

Years ago I heard the same from a Postgres-contractor friend who said amazons ready-to-go (now RDS, but iirc, back then just MySQL; I don’t recall the service brand name - maybe also RDS) SQL service was welcome, because it sucked. Its not clear to me though that 1) it would stay “sucky” 2) users would make the distinction my friend did.

I’m worried it’d play out where Amazon would vacuum up the “low end” part of the market and the “high end” would further have to be competed for, because Joes Expert RDBMS Service v Amazon.

[maxdemarzi]

Sure. Amazon Neptune is garbage, but that's not the issue. They are competing fair and square here. The problem is when they fork <insert liberally licensed open source project> completely.