July 2019 to February 2020​
​

In February 2020, we, the storage team at Scaleway, were released a world first. Personally, it was my second, the first one being the bare metal arm cloud back in 2014.

We released “the first non-AWS S3 compatible GLACIER storage class”. This was a year after the first region in General Availability in Amsterdam. Nine months after the Paris region in General Availability.

​
The wording had been carefully chosen. I remember pushing back on the initial marketing release draft. I pushed back because I felt it was not accurate. When this phrasing came back from marketing, we were all surprised. We checked. Lo and behold, it was true. We were the first non AWS cloud provider to release S3 compatible GLACIER storage class.

The release of the feature was done in two phases. First we had released the lifecycle policy and engine a few months earlier, in June 2019. The work for this had been done by Nicolas (a).

The GLACIER storage class had been testable internally since early summer 2019. But we didn’t want to release it to GA until we had the customer console implementation. Anyhow, it was a bad idea to release it during the french summer holidays.
​
I spent some time digging up how the AWS did their UI and did several meetings with the UX/UI people at Scaleway to advance that track. We used the summer to hammer out the bugs in the backend implementation. Louis (b) did that with some heavy testing and edge case testing to eliminate the last issues. By September 2019, the storage side was ready, we had to wait for the rest to catch up. We didn’t want to trigger lightning again.

The release of the GLACIER storage class was successful. It was so successful that Nicolas quickly had to do a second version of the lifecycle engine. Moving from a single program to a fully distributed program that was able to follow the customer usage curve.

This episode taught me a lesson. Marketing is important. I think as engineers we can sometimes roll our eyes at the marking fluff. At least I do. But I believe that you can do accurate marketing. Understanding what the customer wants, understanding the constraints, the markets and finally boiling down to first principles is paramount. Only then does the distillation make good marketing. But to do this, you have to overcome the initial distaste. You have to put yourself in the shoes of the marketing team. Understand the desired outcome. Find ways to make it work. Find how to collaborate. The same lesson applies to every department. I would learn that later on.

Find something that you do differently and talk about why it matters to the customers or the rest of the ecosystem. That is a good start.

As Scaleway was gearing up for the general availability of the GLACIER storage class, we had a lot of production to handle. The team was stretched. To handle the load we needed to recruit new people in the storage team.

1. You can find a blog post on Scaleway’s website that goes into details: https://www.scaleway.com/en/blog/behind-the-scenes-c14-cold-storage/​
2. ​https://x.com/Scaleway/status/1232957928614899713​
3. ​https://x.com/Scaleway/status/1377290895226241025​

(a) Nicolas Sampré, Storage DevOps at the time, now Staff backend engineer at Qonto

(b) Louis Solofrizzo, Storage DevOps at the time, still at Scaleway

If you have missed it, you can read the previous episode here​

To pair with :

• Gargarin - Public service broadcast, Psychemagik remix
• EBADE