Deploying Cloud Assess

In this tutorial, we assume you are a Cloud Provider who wants to display the environmental footprint of its virtual machine service to its customers.

This tutorial will guide you through

• Initial configuration of Cloud Assess
• Deploying Cloud Assess in a Kubernetes environment
• Automating configuration updates

For this tutorial, you will need to checkout the Cloud Assess repository

git clone https://github.com/kleis-technology/cloud-assess

You will also need

• access to a Kubernetes cluster
• command line tool helm

Configuring Cloud Assess

The folder trusted_library is organized as follows

• The lib folder contains the lca as code models (*.lca files). They encode the allocation rules to compute the environmental impacts of various services (e.g. virtual machines).
• The inventory contains your inventory, i.e. the list of all your physical equipments, as well as their impact factors.

Configuring Cloud Assess consists in filling in the various inventories and their related impact files. We assume that you edit the inventory and impact files in the folder inventory.

Datacenters

For this tutorial, we will assume that you have 3 datacenters located in 3 different countries (Switzerland, Germany, and France). These datacenters are described in dc_inventory.csv. For each datacenter, you need to provide

• id, a unique identifier, e.g. dc-ch, dc-de or dc-fr
• geo, a geographic location, e.g. CH, DE or FR
• total_power, the nominal power (in kW) of the whole datacenter, e.g. 5000 kW
• lifespan, the estimated lifespan (in year) of the datacenter, e.g. 25 year
• reserved_power, the power (in kW) you have reserved in the datacenter, e.g. 100 kW
• power_usage_effectiveness, the power usage effectiveness (dimensionless), e.g. 1.2

Next, you must fill in the impact factors. First, the file dc_impacts.csv must contain,

• for each datacenter identified by its id,
• for each lifecycle step manufacturing, transport, use and end-of-life,
• the impacts for all the environmental indicators.

You must also fill in the impact factors of the electricity mix, in the file electricity_mix.csv. This file contains

• for every geographic location,
• the impacts for all the environmental indicators.

Equipments

Cloud Assess considers three kinds of (physical) equipments

• servers
• storage equipments
• network equipments We focus on the servers in this tutorial. The other kinds of equipments are treated similarly.

The file server_inventory.csv should contain the list of all servers that are necessary to run your service. In our case, this means all servers where the client virtual machines run, and, possibly, other servers (e.g. admin or monitoring).

You can organize your servers in pools. For instance, if you have two virtual machine offers (e.g., basic and performance), corresponding to two disjoint of server resources, you can form two pools of servers.

For each server equipment, in the inventory file server_inventory.csv, you must provide

• id, a unique identifier, e.g. the model reference of the equipment
• pool_id, a unique pool identifier
• dc_id, the datacenter identifier
• power, the average power consumption (in W)
• lifespan, the estimated life span (in year) of the equipment
• quantity, the multiplicity (dimensionless) of this equipment in the pool

Next, you must fill in the impact file server_impacts.csv

• for each server, identified by its unique identifier,
• for each lifecycle step manufacturing, transport, use and end-of-life,
• the impacts for all the environmental indicators.

Deploying Cloud Assess

Packaging the configuration

To package the inventory, you compress the content of the folder inventory in a zip file.

> cd $GIT_ROOT/trusted_library/inventory
> zip -r ../inventory.zip .

Notice that the CSV files must appear at the first level in the archive.

> unzip -l inventory.zip
Archive:  inventory.zip
  Length      Date    Time    Name
---------  ---------- -----   ----
    17153  11-04-2025 00:20   storage_impacts.csv
      156  11-04-2025 00:20   storage_space_inventory.csv
      101  11-04-2025 00:20   dc_inventory.csv
      113  11-04-2025 00:20   compute_inventory.csv
      888  11-10-2025 13:33   network_inventory.csv
      888  11-10-2025 13:31   server_inventory.csv
      950  11-04-2025 00:20   maintenance_impacts.csv
      130  11-04-2025 00:20   vm_inventory.csv
    17153  11-04-2025 00:20   server_impacts.csv
    17153  11-04-2025 00:20   network_impacts.csv
      968  11-04-2025 00:20   dc_impacts.csv
      292  11-04-2025 00:20   electricity_mix.csv
      888  11-10-2025 13:33   storage_inventory.csv
---------                     -------
    56833                     13 files

Deploying in a K8S cluster

First, create a namespace

> kubectl create ns cloud-assess

A Helm chart is available in the Cloud Assess repository under $GIT_ROOT/charts. You can deploy Cloud Assess as follows

> cd $GIT_ROOT/charts
> helm install \
    --set lca.archive.inventory.content="$(base64 $GIT_ROOT/trusted_library/inventory.zip)"\
    -n cloud-assess\
    cloud-assess cloud-assess

Notice that the file inventory.zip is passed to helm as a base64 encoded data. By default, the chart creates a deployment (with 1 replica) and a service. It also creates a configmap containing the inventory.zip content.

Automating configuration updates

If your inventory (and impact files) is updated, you must redeploy Cloud Assess with this inventory. First, package the files under the folder inventory in a zip archive inventory.zip. Then, update the Helm release

> cd $GIT_ROOT/charts
> helm -n cloud-assess\
    update cloud-assess cloud-assess\
    --set lca.archive.inventory.content="$(base64 $GIT_ROOT/trusted_library/inventory.zip)"

The pods should restart (ensuring that Cloud Assess reboots with the new data).

Request an assessment to check that the server runs correctly. Create a port-forward

> kubectl -n cloud-assess port-forward service/cloud-assess 8080:8080

Use curl or any other tool to post a request on localhost:8080. Cf $GIT_ROOT/samples/virtual_machines.http for an example.