FAQ - How to choose an eco-responsible SaaS offer

It is very difficult to be able to compare SaaS software solutions in terms of their ecological sobriety.
However, this should become a criterion of choice like any other when you decide to equip yourself with computer software.
DoliCloud offers you some rational indicators allowing you to compare...

The website

This is probably the least decisive criterion, because, after the visit and the choice of software, the company's portal website will no longer be used. Its footprint is therefore negligible on the rest of the use of the solution. However, it is an interesting indicator (because it is easily measurable and comparable) to see the awareness that the society that offers the solution to the current ecological challenge has.
The site ecoindex.fr allows you to offer an eco-score to a website. The DoliCloud portal site obtains a grade of B or C (variable depending on the state of the site's resource caches, the websites competitors are generally classified from D to G).

The choice of host

Another important element in assessing the eco-responsibility of an online solution is to analyze the hosting provider. It is customary to measure the consumption efficiency of data centers using the PUE (Power Usage Effectiveness). A data center with a PUE of 1.2 consumes 20% more energy than the optimal consumption required to run computing (the perfect PUE being 1). However, this figure is of little interest from an ecological footprint perspective. Here's why:
Imagine a data center in the USA with a perfect PUE of 1 and a data center in France with a less favorable PUE of 1.1. Would the French data center consume, and therefore pollute, 10 percent more than the American one? It would certainly be 10% less efficient, but with French electricity that is 92% carbon-free (without CO2 emissions)! In France, the CO2 emission factor for electrical energy is the best in Europe, with 32 grams of CO₂/kWh (source RTE 2023) compared to 391 grams of CO₂/kWh in the United States (source 2019, it has declined slightly since then). In short, the American localized data center pollutes, for a requirement of 1 MegaWatt of IT processing, at a rate of 1 MW * 1 (PUE) * 391g/kWh = 391 kg of CO2 compared to 1 MW * 1.1 * 32g/kWh = 35.2 kg of CO2. In short, the French data center still pollutes 12 times less than the data center localized in America. In fact, the gap is even greater in favor of French data centers because the PUE is generally better among European hosting providers, which are more sensitive to their carbon footprint.
As you can see, from a data center selection perspective, choosing a hosting provider in France is by far the best choice (the CO2 pollution factor is approximately 13x better; it's not just a question of sovereignty).
At DoliCloud, we have chosen OVH and Scaleway (inter-host replication) .

Greenspector ranked these two hosting providers among the best for their environmental friendliness (with a score of 72 out of 100 each - 2022 ranking). There are other very good ones. And while Greenspector takes into account PUE, it unfortunately does not take into account the CO2 emission factor, which is by far the most important factor (to the point of making the impact of the others negligible). It's regrettable that most rankings of this type make their comparisons assuming an identical electricity emission factor (this is good for comparing efficiency but not absolute pollution, which is all that matters).

But despite all this, this point is not sufficient: Indeed, from one hosting provider to another in the same country, and therefore with the same CO2 emission factor for their electricity, the overall footprint will only vary by a factor of 1 to 3 maximum according to this ranking, whereas from one software program to another, the environmental requirement (imposed by its need for electronic terminals, its need for memory, storage, or CPU, and therefore electronic components, etc.) varies by a factor of 25 (a gain empirically observed through different code and architecture optimization missions on different projects of different sizes).
To simplify, let's say that a well-optimized software program (ecological cost of 1) on A poor hosting provider (cost of 3) will always be much less harmful than a poorly designed, therefore more polluting solution (ecological cost of 25), even hosted with an exemplary hosting provider (cost of 1). Because its footprint will then be (3 x 1 = 3), which is less than the second case (25 x 1 = 25).
The next point will therefore be a criterion just as important as choosing a data center in France.

Optimization of the solution

The software itself and its resource requirements and optimization are undoubtedly the most impactful element in the carbon footprint of an online solution. It is therefore especially important to analyze this footprint which results from the use of the software and the constraints on the electronics and resources used, but also from its manufacturing process. DoliCloud is based on the Dolibarr solution, which is Open Source software. Reusability software components, made possible by Open Source development, but also the robustness and above all the optimization facilitated by the publication of sources, makes DoliCloud a solution much more efficient than most proprietary closed solutions. The choice of technical architecture, made upstream of the development of the Dolibarr software, in order to have a solution with a low memory footprint, CPU and disk widens the gap even further, including with respect to other Open Source solutions. Thus the Dolibarr solution is the only ERP and CRM that can serve an SME, without negative impact on performance, on a simple Raspberry 3 connected to an external SSD drive (still prefer, in production, to use a real dedicated or shared server with an adequate backup policy).
It is therefore important to succeed in quantifying the overall footprint of the life cycle of the use of SaaS solutions in order to compare them with each other...

The measurable footprint of the entire life cycle (manufacturing of the software, hosting of the solution, footprint of the hardware required on the server side, but also user in terms of terminal requirements, needs in electronics, electricity, water, etc.) is undoubtedly the most significant indicator for estimate the eco-responsibility of a solution. Unfortunately, it is also the hardest to evaluate.

We are therefore working to create such a tool (currently in the specifications phase) which will also allow you to compare solutions with each other...
Unfortunately, we do not yet have a date for making this tool available, as the work on such a tool is gigantic and we are looking for partners ready to co-finance this initiative...