Franck Pachot

speaker

138 appearances 1 recordings 1 series first heard Jul 2026 last heard 8 Jul

Franck Pachot’s voice in public audio — every appearance, attributed to the second.

Trend

recordings per month · last 12 months

Recordings per month over the last 12 months — 1 in all, peaking in Jul 2026 with 1.

Appearances

newest first · ▶ plays the moment

Coder Radio · 653: Microsoft's Franck Pachot · 8 Jul 2026 podcast

Hi, I'm good, I'm good, yeah.

I joined Microsoft in the product management team working on Postgres databases on Azure, yeah.

know postgres style data lakes right can you tell me a little bit let's start with horizon db tell me a little about what is that i know it's pretty pretty new yeah yeah yeah it's pretty new because it's in general preview just from this month it was a private preview uh before and it will be generally available in a few months probably so the idea is not uh

completely new the idea is when you run a database in in the cloud you want to scale but sql database is not easy to scale because you need a single point that does the sql the grants the query and and manages transactions so the idea is to separate the compute part and and the storage part

A bit like Aurora did, or AloeDB on Google, Aurora for AWS.

The main reason for that is that it's not complex to scale out the storage, but the transactions part is more difficult.

So the idea is that compute...

is stateless and you have a single instance that manages all transactions there, but the storage can scale out.

So you scale up the compute and you scale out the storage.

So it's it makes scaling easier, especially on the cloud.

Postgres can scale vertically.

It can handle big workloads if you scale vertically.

The problem with scaling vertically is that typically you need to stop the instance and start a larger one.

So you have a downtime.

If you disaggregate those parts, it's easier to scale the compute alone if there is no state in the compute alone.

So the idea is that you scale up only the part that you cannot scale out.

with Postgres without giving up on some features of Postgres, but you scale out the storage there.

So basically the compute instance runs the Postgres code, but plugs some extensions that are specific to the cloud storage.

And also the difference is that it runs on a specific cloud, Azure here, so it can benefit from

features of the storage that may not be available everywhere.

Showing 1–20 of 138 · page 1 of 7