SQLStarters

Platform-as-a-Service databases and the future of the DBA role

I just realized that it has been awhile since my last post, so I figured it’s time to do one before the end of the year. Last few months have been rather hectic with me migrating new workloads to Cloud as well as taking the time to visit PASS Summit 2019 to learn more about Microsoft Data Platform.

I also wanted to finish up my year of blogging by writing about a topic close to my heart, the evolution of the DBA role in the Cloud era. I did have a session about this topic in SQLSaturday Finland earlier this year and I will be doing a PASS Cloud Virtual Group presentation about it on January 16th.

In this post we’ll be looking at one of the Cloud technologies that have significant impact on the DBA role in the future, Platform-as-a-Service databases.

Platform-as-a-Service and the false sense of security

I love the Platform-as-a-Service offerings for the databases in the Cloud. They definitely make our lives easier by removing much of the less exciting toil and freeing our time for more meaningful tasks. However there’s also a false sense of security created by these same services and that easily misleads businesses into believing that in the Platform-as-a-Service world you don’t actually need people with database expertise.

When the businesses start thinking that database experts are not needed, they tend to overlook securing the database competencies. And when something goes wrong, as it inevitably will because databases were left on autopilot, there’s no people in the business with the right skill set to help fix the problems that arose.

This of course works well for consultants who get called in, but maybe not so much for anyone else…

While Platform-as-a-Service databases aren’t replacement to DBAs, they do remove some of the tasks and responsibilities that have historically defined that role. Personally I see this as an opportunity for the DBA role to evolve naturally to meet the requirements of the Cloud era, and to me this is path leading towards Database Reliability Engineering.

I won’t go deeper into that term, but I will explore it more in-depth on my January 16th PASS Cloud Virtual Group presentation.

One of the main reasons I think Cloud is an excellent option for database workloads is, that it can help you lower your operational costs. However with Platform-as-a-Service databases there’s again a little bit of a danger here. Especially when you’re using pay-per-use model, because…

Platform-as-a-Service database performance is made of money

When we were still very deeply rooted into the on-premises world, the idea of throwing money at the performance issue often caused a reaction in most DBAs (and the people who were responsible for IT budgets). And the reaction usually wasn’t utterance of the words “That’s a great idea, lets do it“. It was usually more of a eye rolling and muttering type of thing.

In the Cloud it is, however, exceedingly easy to do just that! A big part of the Cloud promise is that there’s the ability to scale up to almost endlessly and in theory this sounds great, no more performance problems for us. However the thing that’s mentioned less often is that each time you do scale up, the expense of running the databases are scaling up as well.

Let’s look at the example.

The cheap Premium option

This is the cheapest of Premium SQL Databases when you’re using DTU based pricing (the same applies for vCore model too, though). So what happens when the 125 DTUs is not enough and you nudge the money slider (as I like to call it) little bit to the right. Well, you get twice the DTUs and..

Hey! It’s twice the cost too!

Twice the costs. And when you’re again running out of resources and you need that extra boost, it’s again doubling that cost. The more often you need to touch the money slider for increase in computing resources, the quicker you arrive to the point where the lower operational costs that attracted you to the Cloud are just a distant memory.

One of the problems I’ve previously seen is that it’s been difficult to justify the effort of tuning your databases and queries to the business. However with the Platform-as-a-Service databases the actual costs are very much visible and you can impact them directly. Let’s say that you do some query or index tuning to be able to nudge the money slider to left.

Congratulations! You just dropped the operational costs of that database into half of what it was and you can even get the exact amount of money you saved for the business. I personally find Platform-as-a-Service databases to be an exciting opportunity for the database professionals to display what their skills are really worth.

I’d also love to hear that what thoughts you, my readers, have about the future of the DBA role in the future, so feel free to leave comments here about this topic.

And with that I wrap up this post and wish you all happy New Year!

Published by

One response to “Platform-as-a-Service databases and the future of the DBA role”

  1. […] quite a bit, if you allow it, but so will your cloud bill. I’ve previously written about the infamous money slider, which enables you to ramp up the speed of your database, by throwing more money at it. In the past […]

Leave a Reply

%d bloggers like this: