So "build vs. buy" is brought up a lot as a topic on these here internets. The folks with the strongest opinions either seem to have no idea what their time is worth, or have something they're very interested in selling.
An awful lot of engineering types don't value their own time for beans. We start working on computers by puttering around with them when our time is free, and forget that to our employers we cost more than the computers do.
(This is why I went software instead of hardware. When you're accident prone and break software, you can recover from backup. Above a certain threshold, fires are harder to recover from.)
Conversely, folks with something to sell have an intrinsic bias towards not advocating for their own paychecks.
Both ends of that spectrum are largely insufferable.
Both ends of that spectrum are largely insufferable.
The answer is nuanced, and it's not always clear.
Should you use RDS, or run your own databases on top of AWS? They're super opinionated about this despite the fact that they print money either way.
Should you use RDS, or run your own databases on top of AWS? They're super opinionated about this despite the fact that they print money either way.
It's not just your time to build something AND THEN SUPPORT IT. YOU DID REMEMBER SUPPORT BURDEN, RIGHT?
It's the opportunity cost of you building that thing instead of building something that's more aligned with your core business.
It's the opportunity cost of you building that thing instead of building something that's more aligned with your core business.
Conversely, being quoted more than your company's valuation for a product means there's no way in hell someone's going to buy; they're going to build.
This particular point is brought to you eight years ago by @Splunk.
This particular point is brought to you eight years ago by @Splunk.
Let's zoom out a bit to the big question people ask: will they save more money by migrating their data centers to the cloud?
People will weigh in on this as soon as I hit "Tweet," but the honest answer is "it depends."
And your TCO analyses are a nest of lies.
People will weigh in on this as soon as I hit "Tweet," but the honest answer is "it depends."
And your TCO analyses are a nest of lies.
The spectre of lock-in gets raised as justification to avoid buying anything more technically sophisticated than "a wheel," whereas other folks believe their cloud vendors well past the point where it's prudent to do so.
As I said, I'm somewhere in the middle on this.
As I said, I'm somewhere in the middle on this.
You don't need a Cloud Economist to tell you that you should probably delete that unused 4 petabytes of data in S3, or that turning your developer environments off when they're not in use saves you money.
That's easy stuff. The nuanced stuff... carries more weight. :-)
That's easy stuff. The nuanced stuff... carries more weight. :-)
And that's true of most vendor offerings.
You're not going to sit down and beat together a better database over the weekend than a database vendor is going to be able to deliver (once you remove MongoDB from the list).
Same story with logs, monitoring, and a pile of others.
You're not going to sit down and beat together a better database over the weekend than a database vendor is going to be able to deliver (once you remove MongoDB from the list).
Same story with logs, monitoring, and a pile of others.
