- 12/05/2025
Webinar: How to assess the environmental impact of data centers ?
Discover the update of the RCP methodology supported by ADEME to assess the environmental impact of data centers and the cloud.
At a time when there are increasing announcements of data center projects in France and around the world and studies warning of the upcoming explosion of their environmental impacts, we offer a presentation of the RCP Data Centers and Cloud Services.
This methodology based on life cycle analysis (LCA) allows data centers and cloud providers to assess the environmental impact of their services in order to share it with their customers and feed into their eco-design approach.
Webinar speakers:
• Julia Meyer, Digital sobriety engineer, ADEME
• Alexis Perez, Digital sobriety engineer, ADEME
• Louise Aubet, Lab Pole Coordinator, Resilio
PROGRAM
Introduction and issues on data centers
• Main environmental issues on data centers
• Recent and ongoing works on data centers
The RCP Data Centers and Cloud Services
• The objectives of the RCP
• Presentation of the RCP methodology
The webinar took place on Thursday, April 10 from 5 p.m. to 6 p.m.
Discover the update of the RCP methodology supported by ADEME to assess the environmental impact of data centers and the cloud.
At a time when there are increasing announcements of data center projects in France and around the world and studies warning of the upcoming explosion of their environmental impacts, we offer a presentation of the RCP Data Centers and Cloud Services.
This methodology based on life cycle analysis (LCA) allows data centers and cloud providers to assess the environmental impact of their services in order to share it with their customers and feed into their eco-design approach.
Webinar speakers:
• Julia Meyer, Digital sobriety engineer, ADEME
• Alexis Perez, Digital sobriety engineer, ADEME
• Louise Aubet, Lab Pole Coordinator, Resilio
PROGRAM
Introduction and issues on data centers
• Main environmental issues on data centers
• Recent and ongoing works on data centers
The RCP Data Centers and Cloud Services
• The objectives of the RCP
• Presentation of the RCP methodology
The webinar took place on Thursday, April 10 from 5 p.m. to 6 p.m.
Catégorie
🗞
NewsTranscription
00:00Ok, nous allons commencer.
00:05Nous allons commencer par une heure pour présenter le PCR Data Center on Cloud Services.
00:12C'est une méthodologie environnementale de méthodologie
00:18basée sur l'analyse pour évaluer un Data Center hosting on cloud services.
00:23Nous allons avoir environ 45 minutes de présentation et 15 minutes de questions et answers à l'end.
00:32Il y a une section de Q&A sur les thèmes, donc vous pouvez les poser vos questions maintenant et nous allons les lire à l'end.
00:43Je voudrais aussi vous montrer que la session est enregistrée.
00:47Nous allons maintenant passer à l'introduction du speaker.
00:56Je vais donner la parole à Julia et Louise pour se présenter.
01:00Merci Alexis. Hello, tout le monde.
01:04Je suis Julia Meyer de l'ADEME team.
01:08C'est l'agency de l'écologie de transition écologique.
01:12Et je travaille sur l'assessant l'impacte des technologies digitales.
01:19Donc, c'est pour construire PCR.
01:23Et aussi, je travaille sur les études pour mieux comprendre et évaluer l'impacte des technologies digitales.
01:35Louise, peut-être ?
01:38Je ne sais pas si elle est en train de travailler.
01:48Peut-être Alexis, peut-être ?
01:50Alexis, peut-être ?
01:55Oui, pardon.
01:57Je m'appelle Alexis Pérez et je travaille dans la même team que Julia,
02:01à la digital sobriété engineer.
02:04Je suis en train de passer la PCR pour le Data Center sur cloud services,
02:10qui Julia a travaillé sur plusieurs années.
02:14Et je remercie à elle pour tout le travail qu'elle a fait.
02:17Je vais vous entendre.
02:19Je ne parle pas très bien en anglais,
02:23donc je vous remercie pour mon terrible français accent.
02:27Je vais vous remercie pour une autre erreur.
02:29Je vais vous remercie pour une autre erreur.
02:31Je vais vous remercie pour une autre erreur.
02:35Je vais vous remercie pour une introduction avec une présentation de l'impacte.
02:41Ensuite, les principales challenges face aux data centers,
02:45et les récentes et les étudiants.
02:49Je vais vous remercie pour une autre erreur.
02:51Puis, Louise,
02:52je vais vous présenter la PCR pour le Data Center sur cloud services,
02:55avec une présentation de la méthodologie,
02:58avec une exemple de calculations basées sur la PCR.
03:03Enfin, nous aurons une conclusion avec les développement prospects sur la Q&A.
03:10Et maintenant, je vais passer à Julia pour la présentation de l'impacte.
03:17Merci Alexis.
03:18Merci Alexis.
03:20Et Louise, vous avez le temps pour vous présenter.
03:22Juste après.
03:24Donc, oui, quelques mots sur ce que nous faisons sur l'ADEME.
03:28Juste à dire que la méthodologie PCR a été construite
03:35en partie de l'ADEME.
03:37Et nous sommes 14 personnes dans cette équipe.
03:41Nous travaillons sur différents sujets liés aux digitales sobrières.
03:45de l'ADEME.
03:47Donc, c'est de l'ADEME.
03:49de l'ADEME.
03:51C'est de l'ADEME.
03:53de l'ADEME.
03:55de l'ADEME.
03:57de l'ADEME.
03:59de l'ADEME.
04:31de l'ADEME.
04:33de l'ADEME.
04:35de l'ADEME.
04:37de l'ADEME.
04:39de l'ADEME.
04:41de l'ADEME.
04:43de l'ADEME.
04:44de l'ADEME.
04:45de l'ADEME.
04:46de l'ADEME.
04:47de l'ADEME.
04:48de l'ADEME.
04:49de l'ADEME.
04:51de l'ADEME.
04:52Un des studies que nous avons conduit,
04:57que nous avons publié récentement...
05:00Alexis, pouvez-vous switcher la slide, s'il vous plaît ?
05:04Oui, merci.
05:06Une des studies que nous avons conduit récentement...
05:10montre que la technologie digitale est de 4.4%...
05:18de la france de carbone.
05:21Et aussi, c'est de 11%...
05:25de la france de l'électricité.
05:30C'est de 50% de la carbonearbeit...
05:36de l'adjemesse de l'ad derechos des utilisés...
05:39sur notre smartphone, la computer, la tv...
05:44l'outil, etc.
05:45Et 46% sont de l'adolesse de l'adinken.
05:51Alors, c'est de l'administration...
05:55de l'adaptation de l'adaptation...
05:57account for the majority of the carbon footprint but recently in our new study we included in
06:07the perimeter of the evaluation the data centers that were that are abroad to that and use for
06:17French usage so that's why the data center part is more input more important now in the results
06:26and also we can say that this impact is surely underestimated right now because those estimations
06:36didn't take into account the AI part and the computing power associated with those new
06:47technologies I let Alexi go into detail on for the impact for data centers part thanks Julia yes on
07:01data centers the graph on the left shows the impact of data center on the climate change indicator with
07:0965 percent impacting the consumption phase on 35 the use phase on 35 percent impacting the manufacturing
07:18distribution and end-of-life phases but this figure should be taken with caution as it depends greatly
07:27on the local context when considering data centers located abroad the carbon impact of the consumption
07:33phase is greater if we only consider French data centers the manufacturing phase is the most significant
07:40so on the white the manufacturing phase the manufacturing of server on storage equipment accounts
07:50for the majority of the impacts as they are the most numerous pieces of equipment on of the greatest
07:58individual impacts for consumption we have it consumption cooling on the consumption of support equipment
08:07which share the impacts I will now talk about the French on European regulatory context I will briefly go over
08:22the regulatory context because it doesn't necessarily necessarily concern everyone listening to this
08:28webinar but it can still give an idea where of where France on the European Union stand with regard to
08:35obligations so the first is the tertiary eco energy decree this requires an annual decoration for buildings over
08:47under 1000 square meters there are final energy reduction targets from 30 percent in 2030 to 60 percent in 2050 compared to a baseline year
09:04then the SREN law this concerns cloud providers who will have to provide data such as carbon footprint
09:15as water consumption on energy consumption a decree will be issued specifying to the procedure on the
09:24stakeholders involved there are currently discussions for the PCR to be used under this law and there are other
09:33regulations such as the collection of environmental data by RCEP which requires RCEP is the French regulatory authority
09:43for electronic communications so it requires data on energy consumption water consumption on land use for example
09:51there are also also the data center decarbonization roadmap which is a voluntary commitment by industry stakeholders
10:00and the REN law with obligation on reducing waste heat water use etc
10:07now at the European level an important regulation is the energy efficiency directive it divides regulatory obligations
10:21according to the power of the data center above 500 kilowatts there are data provision requirements above one megawatt
10:31but there are more obligation obligations like a requirement to reuse waste heat
10:38now I will describe some recent and ongoing work by ADEM on the subject of data centers
10:49but since this work is not in English I will quickly go over it
10:54first we have the updated ADEM figures we have a major study measuring the environmental impacts of digital technology in France
11:07which we have updated to 2025 Julia has already spoken about this these are the figures presented at the beginning with the environmental impacts of terminals data center on the network
11:23then expert opinions on data centers which provide context regulations on priority actions to limit the impact of data centers
11:38and important work is a prospective prospective study on data center to 2060 to estimate the environmental impact of data centers
11:45to estimate the environmental impact of data centers taking into account AI, cryptocurrency, blockchain, quantum computing, etc.
11:59with development trajectories the work will be completed in September 2025
12:06and finally cooling study which provides an overview of cooling system on emerging technologies
12:14and now I give the floor to Julia for the presentation of the PCR
12:25yes thank you Alexis so this work of PCR is based on the PEF methodology so it's as it was mentioned before
12:37is a methodology based on LCA and the goal is to first inform the consumer or the user of the environmental impact of the service digital service or product that they use
12:56based on the same way of evaluating the impact so to have the same parameter the same
13:04and cutoff rules etc and on the other part is to encourage a concept or manufacturers to do some eco design and so to have a big overview of where is the impact or the greatest impact of their service and to be able to reduce it
13:31so we have already published several PCR so the first that is the parent the digital service PCR and there is also the information system PCR the PCR on internet service provision the LAN network and the one that interest us today data center and cloud services
13:58and so to conduct this work there was a different step so in 2020 the AJEC law was the beginning of the first PCR that was conducted 2021 it was the beginning of the work for data center and cloud services
14:25this work ended on in 2023 and in 2024 we conducted collective action with several participants that did their LCA at the same time using this methodology to be able to be sure that it was complete and we updated
14:54updated with this collective action with this collective action and that's what we present today so thank you again to the participants of this collective action this year
15:10I give the floor to Louise to present the PCR more into detail
15:17if her mic is working because I think she has some issues
15:25on the other hand
15:32ok
15:40ok
15:41J'ai réactivé l'écran, je sais.
15:51Ah oui, ok.
15:53Normalement.
15:54So, we have some mic issues,
15:58but
15:59everyone should be able to
16:05to speak right now.
16:11Oh, it seems to work.
16:24Okay.
16:25Est-ce que vous m'entendez là?
16:26Yes.
16:27Oui.
16:27Okay, sorry.
16:29No problem.
16:31Finally, it works.
16:33Yeah, okay.
16:34Sorry for this technical problem.
16:37I'm very sorry, but yeah, let's continue.
16:43Right on time.
16:44And not waste any more time.
16:48Everything is ending well.
16:51Sorry.
16:52Yes, so I, due to this technical problem,
16:54I could not present myself previously.
16:57So, I'm Louise Obey.
17:01I work at Resilio,
17:03and we participated in this update
17:06of this product category role
17:08on the data center in cloud services
17:11last year.
17:13And that's why I'm here
17:16to present now this document.
17:19So, this new version is available online
17:22on the ADEM bookstore in French and English.
17:27So, here we'll give a brief summary
17:32of the content,
17:33but don't hesitate to go and read it
17:35in more detail.
17:39So, as mentioned before by Alexi and Julia,
17:43so this document is really focusing
17:45on the methodology
17:48and on how to assess the environmental impact
17:51of data centers
17:53and cloud services more specifically.
17:56So, if, I don't know,
17:59in your needs,
18:01then you need to also to get
18:03and to apply a methodology
18:06on other ICT digital services
18:12or other information
18:17and telecommunication technologies,
18:20then, in this case,
18:22other PCR would be relevant
18:26and you can apply this methodology
18:30for this other parts
18:34like network
18:35or end-user equipment, for example.
18:38So, here, we'll only focus
18:40on data center and cloud services.
18:44And you can go to the next slide.
18:47And so, this document is based
18:53on a lifecycle assessment approach.
18:57And with this, in this framework,
19:01we focus on what are called
19:03functional units.
19:05So, these are main units
19:10and blocks on which we are focusing
19:13in the assessment.
19:14There are various ones
19:16in the PCR document.
19:20And so, here, let's have a look
19:23at the various functional units.
19:28They are organized here,
19:31so in this graph,
19:32in kind of layers
19:33and they are all dependent,
19:36the one on the other.
19:37I will come back to that
19:38a little bit later.
19:39So, first, let's start
19:41with the functional unit number one
19:43at the bottom of the slide.
19:46So, provide hosting services
19:48for operation of the IT equipment.
19:50So, this is really the basis
19:53and it concerns building,
19:57so cooling,
19:58also electricity provisioning, etc.
20:02So, everything concerning the services
20:06that you can find around
20:08that would allow the equipment to work
20:11and then the cloud services to run.
20:14And then, let's go to the second layer,
20:17which is bare metal.
20:19We can find,
20:20so, three functional units
20:22on that topic.
20:24So, the functional unit number two
20:27focuses on physical server,
20:30then functional unit number three
20:32focuses on storage equipment
20:35and functional unit number four,
20:39sorry,
20:40focuses on network equipment.
20:43So, here,
20:44they are really the functional units
20:48that are focusing on the hardware
20:50and physical equipment
20:52and they are taking into account
20:56and also impact linked to housing
21:00as they kind of call the impact
21:03based on functional unit number one
21:07concerning housing.
21:10And then,
21:10the third layer is infrastructure
21:13as a service,
21:14yes,
21:15so, it's already digital services.
21:17and functional unit number five
21:21is about processor calculation,
21:25time of processor calculation.
21:28It is very specific
21:29to high performance computing,
21:32for example,
21:33or scientific computing.
21:35So,
21:35it is a bit of a specific use case.
21:40So,
21:41let's focus on functional unit six
21:43and seven,
21:44which are focused on
21:45virtual server
21:47and block storage space.
21:50So,
21:50these are the two main services
21:52that we can find
21:54among cloud,
21:56various cloud services.
21:58And then,
21:59if you want
21:59to build,
22:01if you have
22:02a more custom function,
22:04a more custom cloud service,
22:07then you can create
22:09your own
22:10functional unit
22:12using
22:14the already existing
22:16functional unit
22:17as
22:17Lego bricks,
22:19I would say,
22:20and add them.
22:21And,
22:21of course,
22:22there are examples
22:22on how to do that
22:23in the PCR.
22:28Yeah,
22:28you can go to the next slide.
22:32So,
22:32let's
22:33dive a bit
22:34in more details
22:35and take
22:36the example
22:37of functional unit number two.
22:39So,
22:39provide a physical server
22:41hosted in a data center
22:42with a given
22:43computing capacity.
22:46So,
22:46for each
22:47functional unit,
22:48and in this case
22:49concerning physical server,
22:51the idea
22:53is to
22:53give
22:55information
22:57on the scope.
22:59So,
22:59what should we take
23:00into account
23:01when we want
23:03to assess
23:03the environmental impact
23:05of a physical server?
23:07Then,
23:07for example,
23:10in this case,
23:11we,
23:11of course,
23:11need to take
23:12into account
23:12the dedicated
23:13IT equipment
23:14that we are considering,
23:16the physical server,
23:17but also
23:18the shared
23:19IT equipment
23:20that are needed
23:21for this
23:23specific server
23:25to run
23:26and to do
23:27what it is
23:27meant to do.
23:28So,
23:29for example,
23:29a network,
23:30shared network
23:31equipment,
23:31there can be
23:35a lot of them
23:35in the data center,
23:38but they will be
23:38shared and used
23:39by many other
23:41servers.
23:44And,
23:44of course,
23:45it is important
23:46to explain
23:49and to document
23:51the life cycle phases
23:53that are considered
23:54in the assessment.
23:56And here,
23:56with this
23:57life cycle perspective,
23:59we are taking
24:00into account
24:01the impact
24:03during the whole
24:04life cycle
24:05of the equipment
24:06or the service.
24:07So,
24:08here,
24:08going from the extraction
24:09and manufacturing
24:10of the components
24:12to the end of life,
24:13also taking into account
24:15distribution
24:16and usage,
24:18for example.
24:20And,
24:20last but not least,
24:22all the allocation rules
24:24that are needed.
24:25so the VCP
24:26on how to assess
24:29the environmental impact
24:31or the computation
24:33and calculation rules
24:34are given.
24:36So,
24:37for example,
24:38here,
24:38if we want to assess
24:39the environmental impact
24:41of this specific server
24:42during the use phase,
24:45then there is this
24:47equation here
24:48that is given,
24:50which means that
24:51the impact
24:52will depend
24:53on the duration
24:54of the study
24:56that you consider,
24:57of course.
24:58So,
24:58it can be
24:59one month,
25:00one hour,
25:01as you define it.
25:05It will also depend
25:06on the equipment
25:07power.
25:11So,
25:12it can be the,
25:14so,
25:15yeah,
25:16so it will be
25:16the average power
25:18of this equipment
25:19during the duration
25:23of study
25:23that you consider.
25:24So,
25:24during this one month
25:26that you consider,
25:27for example,
25:27what is the average
25:28power consumed
25:29by this equipment?
25:32And,
25:32last parameter
25:33is the electrical mix.
25:36So,
25:36depending on
25:37the country
25:39or geographical region
25:41in which
25:42the equipment
25:44is located,
25:45then the impact
25:46related to electricity
25:47use will not
25:48be the same.
25:49So,
25:49it's important
25:49to differentiate
25:50that too.
25:52So,
25:52in this case,
25:53the allocation rule
25:55is quite simple,
25:56as you can see,
25:57but in some
26:00more complex
26:01functional units,
26:03for example,
26:03six or seven,
26:04the allocation rules
26:05will be much
26:07more detailed
26:08and complex
26:09as you need
26:10to take into account
26:11many different
26:12equipments
26:14and allocate
26:15the impacts
26:17in a precise way.
26:19And this is all
26:20detailed
26:20in the PCR
26:22cloud and data center.
26:25Yeah,
26:25you can go to
26:26the next slide.
26:27Thank you.
26:28And,
26:28so,
26:31as mentioned,
26:32before,
26:32so there was
26:32a first version
26:33of this document
26:35published a few years ago,
26:36and the goal
26:37last year
26:38was to update it
26:40and to make it
26:42easier to
26:43understand
26:44and to apply,
26:46basically.
26:47So,
26:48we took this
26:49first version
26:50of the document,
26:51applied it
26:52to very
26:53specific
26:56and real case
26:57scenarios,
26:58and then
26:59wrote down
27:01everything
27:02that was
27:02missing
27:03or that was
27:05too complex
27:06and needed
27:06to be simplified,
27:08maybe errors
27:09or inconsistencies
27:10that we found.
27:11And the goal
27:12was to
27:12aggregate
27:14all of these,
27:15all of these
27:17modifications
27:17in order to
27:19obtain
27:19then this
27:20second
27:21version
27:22of the document.
27:23And then,
27:25now we'll
27:26go a little bit
27:27in more details
27:28in some of
27:29these main
27:31modifications
27:31that have been
27:32done to
27:33the document.
27:36So,
27:38one
27:39first
27:40feedback
27:42on this
27:44document
27:44that was
27:45made
27:45and that
27:46was
27:46coming
27:47back
27:48very often
27:49was
27:49the fact
27:50that
27:51there was
27:52a lot
27:53of
27:53imbegweities
27:54and especially
27:55on the
27:56scope
27:56of
27:58assessment.
27:59It's
28:00very
28:00important
28:01to be
28:01clear
28:02on
28:02when you
28:03provide
28:04a
28:04methodological
28:05document
28:05like this
28:06on how
28:07to,
28:08if you
28:09want
28:09the
28:10results
28:11from this
28:13methodology
28:14to be
28:15homogeneous,
28:15to be
28:16able to
28:17compare them
28:17with each
28:18other,
28:18then it's
28:19important
28:19that
28:20we are
28:21all aligned
28:22on what
28:23we are
28:23taking
28:24into
28:24account
28:24and what
28:25we are
28:25not
28:26taking
28:26into
28:26account.
28:27It's
28:27very
28:29important.
28:29So,
28:29a lot
28:30of work
28:30has been
28:30done
28:31on this
28:31document
28:32in order
28:32to
28:33clarify
28:34the scope
28:35of the
28:36functional
28:37units
28:37in this
28:38case.
28:40So,
28:41the
28:41inclusions,
28:43so,
28:43what to
28:44take into
28:44account
28:45for every
28:46functional
28:47unit
28:47has been
28:48listed.
28:49also
28:49the
28:49exclusions,
28:50so,
28:50what to
28:51avoid,
28:52what to
28:53not take
28:53into
28:53account
28:54and
28:54why.
28:57And,
28:58so,
28:59for each
29:00of these
29:01inclusions,
29:03the data
29:04to collect
29:05was
29:07listed
29:08and
29:08all
29:10this data
29:10is then
29:11summarized
29:12in one
29:14big table
29:14that you
29:15can find
29:15in Appendix.
29:19and
29:19second
29:21big change
29:23also
29:24was
29:26the
29:27how to
29:29make the
29:31link
29:31and the
29:32dependencies
29:33in between
29:35the different
29:35functional
29:36units
29:36more
29:37explicit.
29:38So,
29:39as I
29:39showed
29:40before
29:40in the
29:42diagram
29:43with all
29:44the various
29:45functional
29:45units,
29:46then
29:46the goal
29:49was to
29:49make this
29:50document
29:50more
29:51modular.
29:52So,
29:52one
29:52functional
29:53unit
29:53can be
29:56understood
29:56if you
29:58only look
29:59at it,
30:00but you
30:01also need
30:02to take
30:03into account
30:04most of
30:05the time
30:05the results
30:06from other
30:06functional
30:07units.
30:08So,
30:08for example,
30:09here,
30:09let's take
30:10with this
30:11graph the
30:11example of
30:12functional
30:13unit number
30:13seven,
30:14so,
30:15concerning
30:15storage
30:16space.
30:18So,
30:18in order
30:18to assess
30:22the
30:22environmental
30:24impact of
30:25a storage
30:25space,
30:26then you
30:26will
30:27do the
30:29list of
30:30your
30:30equipment
30:31that you
30:31need
30:32to provide
30:33the service,
30:34so,
30:34server,
30:36physical
30:37server,
30:37for example,
30:38storage
30:39equipment,
30:40and network
30:41equipment.
30:41and so,
30:43in order
30:44to assess
30:44the
30:45environmental
30:45impact of
30:46this
30:46equipment,
30:47you will
30:48use
30:48functional
30:49units
30:50two,
30:50three,
30:50and four
30:51that we
30:52described
30:53before.
30:54So,
30:55here,
30:55you can
30:55already see
30:56the link
30:56in between
30:58these functional
31:00units,
31:01but it's
31:01not finished
31:03as these
31:04functional
31:05units,
31:05they also
31:06use the
31:08results from
31:09functional
31:09unit number
31:10one concerning
31:11the houses,
31:12the housing
31:13part,
31:13so,
31:14buildings and
31:14cooling,
31:15et cetera.
31:17So,
31:17it's
31:19really
31:21important
31:22to have
31:24this global
31:25view on
31:26how the
31:26functional
31:27units depend
31:28the one
31:28on the
31:29other,
31:30but we
31:30tried in
31:31the document
31:31to make
31:33it into,
31:36yeah,
31:36to make it
31:36more modular
31:38and so
31:39every
31:40functional
31:40unit can
31:41be understood
31:42on its own.
31:49Also,
31:50one thing
31:51that was
31:52lacking,
31:54actually,
31:54one difficulty
31:55actually of
31:58the first
31:59version of
32:00the methodology
32:02was that
32:03actually
32:04in the
32:05data center
32:06and cloud
32:07domain,
32:09it's sometimes
32:10very hard
32:11to obtain
32:12some of
32:13the data,
32:13especially
32:14for cloud
32:14providers,
32:15they are not
32:15all owning
32:18the walls
32:19and the
32:19buildings
32:20in which
32:20the IT
32:23equipment
32:23are stored
32:25and
32:26installed.
32:26and
32:29therefore,
32:30for them,
32:31in this
32:32case,
32:32it would be
32:33very difficult
32:33to obtain
32:34the data
32:35that is
32:36necessary
32:37for
32:37functional
32:38unit number
32:38one
32:39to assess
32:40the impact
32:41of the
32:41buildings,
32:42of all
32:43the cooling,
32:44all this
32:46technical
32:46environment
32:47about cooling,
32:48electricity,
32:50et cetera.
32:51and
32:53therefore,
32:54we now
32:55provide
32:56generic
32:57data
32:57in the
32:58document,
32:59so it's
32:59not perfect
33:00because it
33:00will not
33:01be
33:01always
33:03adapted
33:04to a
33:04specific
33:05case
33:05if you
33:07have a
33:07very
33:07specific
33:08infrastructure,
33:09for example,
33:10but it
33:11already gives
33:12order of
33:13magnitudes
33:14of the
33:14impact
33:14and they
33:15can be
33:15used
33:16so you
33:17have an
33:18estimation,
33:19an idea
33:20of what
33:21are the
33:23impact
33:23of the
33:24housing
33:25compared
33:25to the
33:26impact
33:27of IT
33:28equipment,
33:28for example.
33:32And
33:32two other
33:35methodological
33:38parts
33:39that were
33:40added to
33:41this document
33:41concerns,
33:43so the
33:43first one,
33:44power
33:44purchase
33:45agreement,
33:46which are
33:46contracts
33:48linked
33:49to
33:50electricity,
33:51so this
33:52really
33:52concerns
33:53the
33:53usage
33:54phase
33:55and
33:55how
33:55to
33:56assess
33:56the
33:56environmental
33:56impact
33:57of the
33:58electricity
33:58that is
33:59consumed.
34:03And
34:03so here,
34:05I will not go
34:06into details
34:07if you want
34:07to know
34:09all the
34:10details
34:11of
34:12this
34:13decision
34:14and the
34:15justifications,
34:16etc.
34:16you can
34:16find
34:17everything
34:17in the
34:18appendix.
34:19But basically,
34:20in the
34:21document,
34:21only the
34:22physical
34:23on-site
34:24power
34:25purchase
34:26agreements
34:26are
34:26taken into
34:28account.
34:30This was,
34:32so this
34:33decision
34:33has been
34:35justified
34:37in
34:38details,
34:38but basically
34:41this was
34:42for kind
34:44of a
34:44trade-off
34:44between
34:45wanting to
34:47have an
34:47incentive
34:48but also
34:49comply to
34:50methodological
34:52rigor
34:54and
34:54in order
34:59to be
34:59homogenous
35:00with also
35:01other
35:01approaches.
35:04and
35:05the
35:06second
35:06methodological
35:09subject
35:09concerned
35:10heat
35:10reuse,
35:11so how
35:11to
35:12take into
35:13account
35:14the fact
35:14that in
35:15some
35:15data
35:16centers
35:17the
35:18heat
35:21that is
35:21produced
35:22by the
35:23IT
35:23equipment
35:24is actually
35:25reused,
35:25it is not
35:26wasted,
35:27it is
35:28reused
35:28for other
35:29uses
35:30and
35:32so
35:33the
35:34addition
35:36to
35:36this
35:36document
35:37on
35:38this
35:38subject
35:41is the
35:42fact
35:42that
35:42besides
35:43the
35:44environmental
35:44impact
35:45of the
35:46functional
35:46unit
35:47in itself
35:48without
35:48taking into
35:49account
35:49heat
35:49reuse,
35:50then
35:50it's
35:51also
35:51possible
35:51to
35:52display
35:53the
35:53environmental
35:53impact
35:54of this
35:55avoided
35:56environmental
35:56impact
35:57actually
35:57associated
36:00with
36:00the
36:01fact
36:01that
36:01heat
36:02is
36:02reused.
36:03And
36:04the
36:05goal
36:05in this
36:06case
36:06was
36:06to
36:07be
36:09an
36:09incentive
36:10in order
36:11to
36:11incentivize
36:16the
36:16actors
36:18of this
36:18domain
36:19to
36:19go
36:20towards
36:20this
36:21kind
36:21of
36:22application.
36:25and
36:29last
36:32element
36:33that
36:33is
36:33very
36:33important
36:34also
36:34to
36:35remind
36:36and
36:36to
36:37be
36:38aware
36:39of
36:39is the
36:40fact
36:40that
36:40they
36:40are
36:40kind
36:41of
36:41two
36:41temporal
36:42approaches
36:45that
36:46you can
36:47have
36:47on the
36:47results.
36:48so
36:48let me
36:50explain.
36:50So
36:51basically
36:53you can
36:54depending
36:55on what
36:56you want
36:57to do
36:58with
36:58the
36:59results
37:00of this
37:01calculation
37:02afterwards,
37:04if you
37:05want,
37:05for example,
37:06to
37:07get the
37:09average
37:09impact
37:10of your
37:10infrastructure
37:11or services
37:12over one
37:13year for
37:13an annual
37:15report or
37:15for extra
37:16financial reports,
37:17et cetera,
37:18then an
37:19average
37:20approach
37:20would be
37:21most
37:21suitable.
37:23So you
37:23need,
37:24in this
37:24case,
37:24to have
37:25data that
37:26is
37:27averaged
37:30over one
37:30year in
37:31order to
37:31take into
37:32account the
37:33seasonal
37:34variations,
37:35for example.
37:36But on
37:37the other
37:37hand,
37:38if you
37:38want to
37:38have more
37:39detailed
37:40results
37:41every
37:42hour,
37:43every
37:43day,
37:43for example,
37:44to show
37:46the variation
37:46of the
37:47impact
37:47with
37:48time,
37:49then in
37:50this time,
37:50a more
37:50instantaneous
37:51approach
37:52would be
37:53relevant.
37:56And in
37:56this time,
37:57you don't
37:58need to
37:58collect
37:58average
37:59data,
37:59but rather
38:00collect
38:02instantaneous
38:04data,
38:05but update
38:06it very
38:08often,
38:08every day
38:09or every
38:09hour,
38:11depending
38:11on the
38:12frequency
38:14that you
38:14choose.
38:15So these
38:16two
38:16approach
38:17all
38:17possible
38:18with this
38:19methodology,
38:21with this
38:21document,
38:22but there
38:22are some
38:23constraints
38:24on the
38:26data
38:26collections
38:26that will
38:28not be
38:28the same.
38:29And so
38:30there is a
38:32whole section
38:32in this
38:33document that
38:33details,
38:34depending on
38:34this approach,
38:36what are the
38:37specificities
38:38of the
38:39data that
38:39you need
38:40to collect
38:40depending on
38:41this
38:42average
38:43subject
38:43or not.
38:46And
38:46maybe last
38:48element
38:48concerning
38:49this
38:49temporal
38:50representativeness
38:52is the
38:53fact that
38:54there is
38:54one
38:54specificity
38:55also in
38:55this
38:56domain,
38:57is this
38:57ramp-up
38:58period.
38:59It means
38:59that at
39:00the beginning
39:01when you
39:01open or
39:03when you
39:04install a
39:04new
39:05service,
39:05a new
39:06offer,
39:13then there
39:14might be
39:14a ramp-up
39:17period during
39:17which you
39:18might have,
39:19for example,
39:20just a few
39:21clients on
39:22the whole
39:23infrastructure.
39:26And that
39:27might lead
39:28to, at
39:29the beginning,
39:30really huge
39:30impact,
39:31for example,
39:32associated
39:32to two
39:34clients because
39:35your
39:35infrastructure
39:36is at
39:37the beginning
39:37only used
39:39by two
39:39clients.
39:40And then
39:40after six
39:41months,
39:41one year,
39:42it reaches
39:43a steady
39:44state and
39:44then the
39:45results are
39:46more,
39:46not
39:49coherent,
39:50but more
39:51average and
39:52more steady.
39:54And so
39:55for this,
39:56this kind
39:57of leads
39:57to
39:58some
40:01specificities
40:02in the
40:03communication
40:04format.
40:04so in
40:05this case,
40:05we also
40:06precise some
40:08rules and
40:09constraints on
40:10how to find
40:12the trade-off
40:14in between
40:14this specific
40:16period that is
40:17not representative
40:18of really
40:20the cloud
40:21services on
40:22the stationary
40:23period,
40:26but not
40:28fall into
40:29greenwashing
40:31communication.
40:33So there
40:34is also
40:35a specific
40:36section in
40:37this case
40:38that will
40:39give a little
40:41bit more
40:42details about
40:42that.
40:43and about
40:50this
40:52methodology,
40:53so it's
40:54also,
40:55it's often
40:57some questions
40:57that are
40:59made about
41:00what is the
41:01alignment of
41:02this methodology
41:03compared with
41:04other standards.
41:05so the
41:08alignment
41:10analysis
41:10between
41:11this
41:11document
41:12and the
41:12GAG
41:12protocol
41:13has already
41:14been made
41:15in the first
41:15version,
41:17and you have
41:17some elements
41:18that are listed
41:19here on the
41:20slide,
41:20and what was
41:22added in
41:22this second
41:23version was
41:24kind of
41:26an overview
41:27of the
41:28alignment
41:28of this
41:29document
41:29with the
41:31other
41:31main
41:32standards
41:33that concern
41:34data centers,
41:36cloud services,
41:38et cetera.
41:39So here,
41:40for example,
41:41on this
41:41graph,
41:43you can see
41:44all the
41:45main
41:45standards
41:46about data
41:47centers and
41:47cloud,
41:48depending on
41:50the environmental
41:51criteria that
41:52are considered,
41:54so horizontally
41:56and vertically,
41:57depending on
41:58the scope
41:59that is
42:00considered,
42:00so from
42:01the data
42:01center to
42:02rather the
42:04service
42:06point of
42:07view,
42:08the cloud
42:08service
42:08point of
42:09view.
42:12And
42:12so,
42:15okay,
42:16so now
42:16that we've
42:17seen the
42:18main
42:18additions
42:20and modifications
42:21of this
42:22methodology,
42:23let's take
42:24a few minutes,
42:25I'll try to
42:26be brief,
42:26so we have
42:27time for
42:27questions then,
42:28with one
42:29example of
42:31computation,
42:32so the
42:34idea is
42:36now that
42:37you have
42:39this methodology,
42:40how to
42:40apply it
42:41like in
42:42a real
42:43case.
42:45So one
42:46thing that
42:48would be
42:48super useful
42:49because as
42:50you see,
42:51there are
42:51many functional
42:53units that
42:54are dependent
42:55the one
42:56on the
42:56other,
42:56so doing
42:57the calculation
42:58by hand
42:59can be
43:00quite long.
43:02So in
43:02this case,
43:03it is
43:04rather
43:05more efficient
43:07to use
43:07a calculator,
43:08so in our
43:09case,
43:09we use
43:10the open
43:10source tool
43:11CloudSS,
43:12but it's
43:13also possible
43:14to use
43:14any LCA
43:15software
43:16or a home
43:16calculator
43:17that you
43:17would have
43:18built
43:19that is
43:21using
43:22the
43:23PCR
43:25Cloud and
43:26Data Center
43:27rule,
43:27so all the
43:28allocations that
43:29I was
43:29mentioning
43:30before,
43:31and then
43:32you need to
43:33feed this
43:34calculator,
43:35this model
43:35with
43:36first
43:38usage data,
43:40so for
43:40example,
43:40the number
43:41of vCPUs
43:42of the
43:43virtual
43:44machines that
43:45you have
43:46reserved,
43:46the number
43:47of the
43:47amount of
43:48RAM or
43:49storage also
43:50that is
43:50reserved,
43:52so all
43:52this usage
43:53data,
43:54depending on
43:54time,
43:55depending on
43:55the client,
43:56etc.
43:58You also
44:00need to
44:01give the
44:03data on
44:04infrastructure,
44:06so all
44:06the hardware
44:07or infrastructure
44:09data,
44:10so concerning
44:10the equipment
44:11and all
44:12this technical
44:14configuration of
44:15the equipment,
44:16but if we
44:16are considering
44:17functional limit
44:18number one
44:19for example,
44:19concerning the
44:22technical
44:23environment
44:24equipment,
44:25so the
44:25power generator
44:26or batteries,
44:27etc.
44:29And finally,
44:30we need
44:31environmental
44:32data,
44:33so what
44:34is the
44:35impact in
44:36terms of
44:37greenhouse
44:37gases
44:38emission,
44:39or in
44:40terms of
44:40water use
44:41or other,
44:42what is the
44:43impact associated,
44:45what is the
44:46impact of one
44:47specific equipment.
44:49And then,
44:50once you have
44:50gathered all
44:52this data,
44:53the calculation
44:54is made
44:56using this
44:58allocation
44:59from the
44:59PCR,
45:00and then,
45:01in the end,
45:02you get
45:02the footprint,
45:05the environmental
45:05footprint of the
45:06cloud services
45:07that you are
45:08interested in,
45:09in order to
45:12apply it
45:12in whatever
45:13your goal
45:15or deliverable.
45:19And on the
45:19next slide,
45:21on the next slide,
45:22sorry,
45:23we have
45:23one specific
45:26example to make
45:26it very
45:27concrete.
45:30So,
45:30for example,
45:32let's take a very
45:33simple inventory
45:34with only one
45:35server.
45:35we even
45:36don't consider
45:38any shared
45:40equipment here
45:41to make it
45:42very simple.
45:45So,
45:45we consider
45:46one server,
45:47for example,
45:48only one
45:48unit,
45:49one unit
45:52tall with
45:53some specific
45:55requirements,
45:56some specific
45:58technical
45:58configuration
45:59story.
46:00I will not
46:00go through
46:01everything in
46:02detail,
46:02but you have
46:03Sodi idea.
46:05So,
46:06this is
46:07the
46:07equipment.
46:11Then,
46:11this equipment
46:13has an
46:14environmental
46:15impact of,
46:17so,
46:17concerning
46:17the
46:18global warming
46:21potential.
46:22So,
46:23here,
46:23400 kilograms
46:25CO2 equivalent.
46:27This is
46:27a very
46:28example.
46:30And
46:30if we want
46:32to apply
46:33the allocation
46:34as it
46:35was shown
46:36before.
46:38So,
46:39here,
46:39we only
46:40take the
46:41example of
46:42manufacturing.
46:44So,
46:44let's take
46:45the impact
46:46of the
46:47server over
46:48one month.
46:49We do
46:49the computation
46:50applying the
46:52allocation of
46:53the PCR
46:54cloud and
46:54data center
46:55and get
46:56a few
46:57kilogram
46:57CO2 equivalent
46:59for the
47:01dedicated
47:03server.
47:03So,
47:04and then
47:05if we
47:05want to
47:06account
47:06for the
47:07building
47:08and housing
47:09part of
47:11the impact,
47:13then we
47:13would use,
47:14in this case,
47:14functional unit
47:15number one.
47:16So,
47:16this is
47:16the second
47:17formula.
47:18And
47:18also,
47:20applying
47:20this
47:21temporal
47:22allocation
47:23and using
47:24the data.
47:26So,
47:26here,
47:27this
47:2791
47:28is
47:29this
47:31generic data
47:32that I
47:32was mentioning
47:33before.
47:34So,
47:34this 91
47:35kilogram
47:36CO2 equivalent
47:38that is provided
47:40in the document,
47:41then we
47:41get also,
47:43so,
47:44here,
47:44for the
47:44example,
47:453.7
47:47kilogram
47:48of CO2 equivalent.
47:51and the
47:53idea is
47:54the same
47:54for all
47:56the functional
47:56units.
47:58The thing
47:58is,
47:58so,
47:58here,
47:59so,
48:00it is
48:00quite
48:00understandable
48:02and it
48:03can be
48:05explained
48:05on one
48:06slide.
48:06We took
48:07one simple
48:08example,
48:09but you
48:09can do
48:09the same
48:10for many,
48:11many
48:11various
48:12equipment
48:13and for
48:13all
48:14the
48:14functional
48:15units.
48:16So,
48:21and maybe
48:23I hand
48:24over to
48:25you,
48:25Alexis.
48:26Yes.
48:27Thank you,
48:29Louise.
48:30Now,
48:30I will
48:31talk
48:31about
48:32the
48:32next
48:32steps.
48:34So,
48:35we have
48:35a global
48:35development
48:36prospect
48:36for the
48:37PCR
48:38data
48:38center
48:38on
48:39cloud
48:39services.
48:40First,
48:42regarding
48:43the
48:43implementation
48:45of an
48:45eco-label
48:46for
48:47cloud
48:47services
48:48by
48:48TCO,
48:49which
48:50is
48:50based
48:51on
48:51PCR.
48:53I'd
48:53like
48:54to
48:55take
48:55this
48:55opportunity
48:56to
48:56mention
48:56that
48:57there
48:57is
48:57a
48:58presentation
48:58webinar
48:59on
48:59the
49:00subject
49:00on
49:01April
49:0114th
49:02at
49:032
49:03PM.
49:04If
49:04you
49:04go to
49:05the
49:05Sustainable
49:06Digital
49:06Infrastructure
49:07Alliance
49:08website
49:09and find
49:10the
49:10webinar
49:11under
49:11events.
49:13There
49:14are
49:14also
49:15discussions
49:16between
49:16Resilio
49:17and
49:18the
49:18ISO
49:20the
49:21International
49:21Organization
49:22for
49:23Standardization
49:23and
49:24the
49:25ITU
49:26the
49:27International
49:27Telecommunication
49:29Union.
49:30For now,
49:31these are
49:31discussions,
49:32so I
49:33have
49:33nothing
49:34more to
49:35add
49:35about it.
49:38Then
49:39we have
49:40the issue
49:41of adopting
49:42the
49:42PCR.
49:42PDF
49:44document
49:44may
49:45be
49:45a
49:47little
49:47difficult
49:48to
49:48understand
49:49for
49:49non-specialists,
49:51which
49:51is why
49:52we will
49:53be
49:53providing
49:53a
49:55calculator,
49:56a
49:57collection
49:57on
49:58sharing
49:58of
49:58feedback
49:59on
50:00web pages
50:02that
50:03are
50:03more
50:03concise
50:04than
50:05the
50:05PDF.
50:07These
50:07documents
50:08will be
50:09produced
50:09in the
50:10coming weeks
50:10or
50:11months.
50:11There
50:11are
50:12no
50:12final
50:12dates
50:13yet.
50:16And
50:16with
50:16this
50:17word,
50:17I
50:17conclude
50:18the
50:18presentation.
50:19It's
50:19now
50:20time
50:20for
50:21questions,
50:22so if
50:23you
50:23have
50:23any,
50:24please
50:24don't
50:24hesitate.
50:25I
50:25will
50:25read
50:27the
50:28questions
50:28asked
50:29so
50:29far.
50:30So
50:31we
50:33have
50:33some
50:33questions.
50:37Yes,
50:38first,
50:38I read
50:38a comment
50:39that
50:39said
50:40the
50:40link
50:40to the
50:41PC
50:42air
50:42was
50:42not
50:43working.
50:45Unfortunately,
50:46there are
50:46technical
50:46problems
50:48with
50:48the
50:49ADM
50:49library.
50:50I don't
50:51know
50:51when
50:52it
50:52will
50:52be
50:52restored,
50:53but
50:54I
50:55confirm
50:55that
50:55the
50:56document
50:56is
50:56online.
50:59So
51:00the
51:00question,
51:01the
51:01first
51:01question
51:01is
51:02what's
51:02mint
51:03by
51:04technical
51:05slots?
51:07It's
51:08so
51:09Julia
51:10and
51:10with
51:10you
51:10can
51:11complete
51:13my
51:13answer
51:14if
51:14you
51:14want.
51:14It's
51:15all about
51:15running
51:17the
51:17building's
51:18infrastructure.
51:19It's
51:19mainly
51:20cooling.
51:21thing.
51:24OK.
51:26Then the
51:28next
51:28question,
51:29how do we
51:31know how much
51:32electricity which
51:33data center consumes?
51:34if it's a question
51:38of data center
51:39consumption from
51:40the ADM
51:42RCEP
51:42study,
51:43we know
51:44this from
51:45inventories
51:46and data
51:47that we
51:48collect
51:48as part
51:50of the
51:51ADM
51:52RCEP
51:52study.
51:52and then
51:59there is
51:59a question
51:59about
52:00the
52:00CSRD.
52:03Julia
52:04has already
52:05answered
52:06this question.
52:07I don't know
52:08if you want
52:08to tell
52:10this
52:10online
52:11or no.
52:14Yes,
52:14just to mention
52:15that
52:15this
52:16PCR
52:17is
52:17of course
52:19linked
52:19to the
52:19CSRD
52:20that
52:22provides
52:23a
52:24framework
52:25to
52:26and to
52:28say how to
52:28publish
52:29the
52:29CSR data.
52:31So
52:31for
52:32climate change
52:33you have to
52:34provide
52:35your impact
52:36on the
52:37scope 1,
52:382 and
52:393.
52:40So
52:40for
52:41organizations
52:41that need
52:42to evaluate
52:43their
52:43scope 3
52:44the
52:46PCR
52:47is a
52:47way to
52:48harmonize
52:48the
52:49way
52:50of
52:50evaluating
52:52the
52:53hosting
52:54part
52:54of
52:55it
52:56and
52:57it
52:57goes
52:58even
52:58further
52:59because
52:59it
53:00provides
53:00also
53:00other
53:01indicators
53:02than
53:02the
53:03impact
53:03of
53:03climate
53:04change.
53:04But
53:05it's
53:05of course
53:05a
53:06link.
53:08Thank you.
53:10There is
53:11another
53:11question.
53:12In
53:13the
53:13allocation
53:14rules
53:14formula,
53:15how is
53:15the
53:16production
53:16of the
53:17equipment
53:17is
53:18taken
53:18into
53:18account
53:19or
53:19is
53:20equipment
53:20production
53:21not
53:21relevant
53:21in
53:22this
53:22case?
53:23I don't know
53:24Julia,
53:24if you
53:25want to
53:26tell your
53:28answer.
53:29Just to
53:30mention that
53:31the
53:32production
53:33manufacturing
53:34part of
53:35the
53:35equipment
53:35is
53:36of course
53:36taken
53:37into
53:37account
53:37and
53:39usually
53:40like
53:41it's
53:41done
53:42usually
53:42for
53:43LCA
53:44is
53:45just
53:45amortized
53:47on the
53:48lifetime
53:48of the
53:51equipment.
53:52So for
53:52example,
53:53for a
53:53server,
53:53it's
53:54about
53:54five
53:55years.
54:00Then
54:01are
54:01impacts
54:02given
54:02on
54:03other
54:03indicators,
54:04human
54:06toxicity
54:06potential,
54:07for
54:07example,
54:08or
54:08just
54:09the
54:09five
54:09mandatory
54:10ones
54:10that
54:11were
54:11shown?
54:14Same,
54:14Julia,
54:15if you
54:15want to
54:15discuss
54:16your
54:16answer,
54:17feel free.
54:18Yes,
54:19I tried
54:19to
54:20answer.
54:22So
54:22it's
54:23PEF,
54:24we
54:25recommend
54:26as the
54:27PEF
54:27methodology
54:28to use
54:28the
54:2916
54:29indicators,
54:30but
54:31for
54:31all
54:31the
54:32methodology
54:32for
54:33digital
54:34technology,
54:35we
54:35give
54:38the
54:38result
54:39for
54:39the
54:40impact
54:40factor
54:41for
54:41only
54:42five
54:42indicators
54:43that
54:43are
54:44more
54:45contributive
54:47to
54:48the
54:48result.
54:51So in
54:52the
54:52methodology,
54:53we give
54:53the
54:54impact
54:55factor
54:55for
54:56only
54:56those
54:57five
54:57major
54:59indicators,
54:59but
55:00in
55:01the
55:01ADEM
55:02database,
55:03you
55:04have
55:04usually
55:05the
55:05factor
55:05impact
55:06for
55:06the
55:0616.
55:11Thank you,
55:13Julia.
55:13Next
55:14question,
55:15why are
55:15only
55:16local
55:17physical
55:17PPA
55:18considered
55:19on a
55:21network
55:21well
55:21interconnected
55:22in
55:23Europe,
55:25even if
55:26the
55:26production
55:26is
55:27far,
55:27as
55:27long as
55:28temporality
55:29between
55:29production
55:30and
55:30consumption
55:30is
55:31answered,
55:32it's
55:32possible
55:33to
55:33take
55:34benefit
55:34from
55:34renewable
55:36electricity
55:40purchases.
55:41Sorry
55:41for the
55:42accent.
55:45PPA
55:45is a
55:46big
55:46question.
55:51Yes,
55:51the problem
55:51with
55:52PPA,
55:54which
55:55are not
55:55on-site
55:56PPA,
55:57is
55:58they
55:58require
56:00guarantees
56:02of
56:02origin
56:03to
56:04measure
56:04them
56:05and
56:06there
56:06is
56:06no
56:06temporal
56:08and
56:09geographical
56:09representativeness
56:11about it.
56:14So,
56:14it was
56:15one of
56:16the
56:17problems.
56:19And
56:19Julia,
56:21you
56:22say
56:22another
56:23thing
56:23in the
56:24answer
56:24if
56:24you
56:24want
56:24to
56:26discuss
56:27about it.
56:27Yes,
56:27same as
56:29you,
56:29but it
56:29was a
56:30big
56:30subject
56:31during
56:32this
56:34long
56:34work,
56:34so I
56:35just invite
56:36you to
56:37look at
56:37the
56:38appendix
56:39E
56:39that
56:40details,
56:41that goes
56:42really into
56:42details
56:43on the
56:44different
56:44type
56:45of
56:45PPA
56:46and
56:47why
56:47we
56:49considered
56:50at the
56:51end
56:51that
56:51it's
56:52relevant
56:52to
56:52only
56:53take
56:53into
56:53account
56:53the
56:54physical
56:54PPA
56:55in the
56:56methodology.
56:56Okay,
57:00thank you,
57:01Julia.
57:01And I
57:02think it's
57:02the last
57:03question.
57:04Are
57:05your
57:05results
57:06available
57:07in
57:07open
57:08source
57:08with
57:08the
57:09source
57:09of
57:09the
57:10data?
57:10Which
57:11LCA
57:11software
57:12did
57:12you
57:12use?
57:13Which
57:13database?
57:18I don't
57:19know if
57:19Louise
57:20and Julia
57:20you want
57:21to
57:21answer
57:23this
57:23question.
57:24Yeah,
57:25maybe.
57:26so the
57:27document
57:28that was
57:29presented
57:29is a
57:30methodology
57:31document
57:32so it's
57:33not
57:34a
57:34quantitative
57:37result.
57:38The
57:38idea
57:38is to
57:39work
57:41on
57:42this
57:42methodology
57:43to
57:43improve
57:44it.
57:46Doing
57:46so,
57:47we
57:47apply
57:48the
57:48methodology
57:48to
57:49specific
57:50real
57:54cases
57:56cases
57:57but the
57:59results
58:00are not
58:00open
58:01source
58:01because
58:01they
58:02were
58:02applied
58:02on
58:03various
58:03cloud
58:04providers
58:05and so
58:06the
58:06results
58:07are
58:08only
58:09for
58:10them
58:10to
58:12keep.
58:13However,
58:14the data
58:15that you
58:15can find
58:16the
58:16generic
58:17data
58:18there are
58:19some
58:20generic
58:20data
58:20that you
58:21can find
58:21in the
58:22document
58:22so
58:22concerning
58:23the
58:24impact
58:24for
58:25functional
58:25limit
58:25number
58:26one
58:26for
58:26example
58:26but
58:27others
58:27that I
58:29mentioned
58:29and
58:30for them
58:31you can
58:31find the
58:32data
58:32in the
58:34appendix
58:34of the
58:35document
58:35there
58:37should be
58:38all the
58:39hypothesis
58:39that were
58:40used
58:41and data
58:42from other
58:43database
58:44if
58:44some are
58:45used
58:46everything
58:46is
58:47explained
58:48and
58:50you
58:51in the
58:53upcoming
58:53weeks
58:55or months
58:56I don't
58:56know exactly
58:57but this
58:57data
58:58should also
58:59be available
59:00in the
59:01near future
59:02in the
59:03EMPRINT
59:03database
59:05to
59:05with all
59:06the
59:07documentation
59:07and if
59:12you want
59:12Julia or
59:13Alexi
59:14to
59:14complete the
59:17answer
59:17but
59:17that's
59:18good
59:19perfect
59:20okay
59:22I think
59:23we're done
59:23with the
59:24questions
59:24so
59:27I want
59:28to
59:28thank you
59:29for
59:29questions
59:30Andrew
59:31and
59:32Louise
59:32thank you
59:33for the
59:33answer
59:33thank you
59:35for
59:36watching
59:36the
59:36webinar
59:37I hope
59:38it was
59:38informative
59:38have a
59:40nice day
59:40thank you
59:43have a
59:46good day
Recommandations
1:00:35
|
À suivre
44:43
1:45:18
51:17
0:39