We have intimate knowledge of the systems we run. We don’t want to go through L1 or L2 for issues we already qualified; conversation should start at L3, possibly with the teams building the platform.

5-year comprehensive support is a strict minimum. 7 to 9-year is our target. Spare parts availability, but also software support: new features, bugfixes, security patches…

Our Infrastructure teams are greatly outnumbered by the quantity of assets they operate. Manual operations should therefore be avoided at all cost to work within our Infra-as-code approach, usually by means of API.

Open source is key for troubleshooting, extended support and security. We aim at working only with companies that offer and support open source software so that we can audit, modify, deploy and maintain autonomously.

We are very keen on sharing what works and what doesn’t, so that things get better for everyone. If ideas and processes can help similar actors, we are happy to share mostly everything.

We are more interested in data sheets than marketing presentations. We thoroughly examine exhaustive (possibly NDA-level) specifications, ask questions, benchmark components & features on our use cases.

Our planning and investments are months or even years ahead, so any valuable piece of information is a potential game changer. The ultimate goal being to influence vendors' designs and roadmaps.

We scrutinize many criteria beyond purchasing value (eg. power consumption, maintenance costs) and think in terms of TCO. This could mean buying more expensive assets if this helps us achieve our mission.

Given similar products, environmental impact would be a strong tie-breaker. Existing and future endeavors to improve impact should be communicated transparently and backed by relevant figures.