<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Antoine Lagier on Medium]]></title>
        <description><![CDATA[Stories by Antoine Lagier on Medium]]></description>
        <link>https://medium.com/@a.lagier?source=rss-65b5dacc42a6------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*JBt4Jkl8NykFwMce85HRdw.png</url>
            <title>Stories by Antoine Lagier on Medium</title>
            <link>https://medium.com/@a.lagier?source=rss-65b5dacc42a6------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Sat, 21 Nov 2026 23:53:47 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@a.lagier/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[A propos du EU Sovereign Cloud de AWS]]></title>
            <description><![CDATA[<div class="medium-feed-item"><p class="medium-feed-image"><a href="https://medium.com/@a.lagier/a-propos-du-eu-sovereign-cloud-de-aws-dbd6a309e752?source=rss-65b5dacc42a6------2"><img src="https://cdn-images-1.medium.com/max/1456/1*ozoenddJgw9-z3k6SdfYMA.png" width="1456"></a></p><p class="medium-feed-snippet">AWS a enfin officialis&#xE9; le lancement de son &#xE9;trange objet &#x201C;AWS EU Sovereign Cloud&#x201D; le 15 janvier 2026.</p><p class="medium-feed-link"><a href="https://medium.com/@a.lagier/a-propos-du-eu-sovereign-cloud-de-aws-dbd6a309e752?source=rss-65b5dacc42a6------2">Continue reading on Medium »</a></p></div>]]></description>
            <link>https://medium.com/@a.lagier/a-propos-du-eu-sovereign-cloud-de-aws-dbd6a309e752?source=rss-65b5dacc42a6------2</link>
            <guid isPermaLink="false">https://medium.com/p/dbd6a309e752</guid>
            <category><![CDATA[souveraineté-numérique]]></category>
            <dc:creator><![CDATA[Antoine Lagier]]></dc:creator>
            <pubDate>Fri, 16 Jan 2026 11:20:33 GMT</pubDate>
            <atom:updated>2026-01-16T11:20:33.218Z</atom:updated>
        </item>
        <item>
            <title><![CDATA[History of Law Enforcement Information Requests at Amazon and AWS]]></title>
            <link>https://medium.com/@a.lagier/history-of-law-enforcement-information-requests-at-amazon-and-aws-c0fcb478117b?source=rss-65b5dacc42a6------2</link>
            <guid isPermaLink="false">https://medium.com/p/c0fcb478117b</guid>
            <category><![CDATA[cloud-computing]]></category>
            <category><![CDATA[sovereignty]]></category>
            <category><![CDATA[cloud]]></category>
            <category><![CDATA[afi]]></category>
            <category><![CDATA[aws]]></category>
            <dc:creator><![CDATA[Antoine Lagier]]></dc:creator>
            <pubDate>Mon, 29 Jan 2024 13:32:31 GMT</pubDate>
            <atom:updated>2024-02-12T19:07:17.110Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*TmMYS2g_gfhNkNNr0P1xIg.png" /></figure><h3>History of “Law Enforcement Information Requests” at Amazon and AWS</h3><p>Failing to agree on an in-depth reform of the FISA law, Joe Biden signed the <a href="https://mashable.com/article/biden-congress-renews-foreign-intelligence-surveillance-act">National Defense Authorization Act</a> on December 22, 2023.</p><p>This law, which has gone virtually unnoticed in Europe, not only extends the U.S. agencies’ prerogatives to intercept data without judicial review until April 2024, but also seems to widen its scope a little further: it extends the companies subject to this obligation from Cloud providers to equipment manufacturers (potentially Cisco…) and hosting facilities providers (potentially Equinix or <a href="https://twitter.com/platombe/status/1751981447693349055?s=61&amp;t=yk0Xsdvi3wKVXOLe_oBEgA">Data4</a>…).</p><p>While this issue has given rise to a wide-ranging <a href="https://www.brennancenter.org/our-work/research-reports/coalition-letter-urges-congressional-leaders-keep-reauthorization-section">public debate</a> in the U.S. (with <a href="https://www.bloomberg.com/news/articles/2023-03-22/apple-google-meta-push-to-curb-us-spying-on-texts-emails">GAFAM officially taking a stand against it</a>), this has not been the case in Europe, where in France only MP Philippe Latombe <a href="https://twitter.com/platombe/status/1738204274285334805?ref_src=twsrc%5Etfw%7Ctwcamp%5Etweetembed%7Ctwterm%5E1738204274285334805%7Ctwgr%5E78d7677ff2f93cd771ce27bcd9b355501370c992%7Ctwcon%5Es1_c10&amp;ref_url=https%3A%2F%2Fwww.01net.com%2Factualites%2Fphilippe-latombe-la-loi-fisa-cest-une-bombinette-qui-pourrait-faire-voler-en-eclat-notre-souverainete-numerique.html">has raised the subject</a>.</p><p><a href="https://mashable.com/article/biden-congress-renews-foreign-intelligence-surveillance-act">Congress extends controversial government spy program for 2024</a></p><p>This is the latest in a long series of news items on the subject of data sovereignty, including the successive invalidation of Safe Harbor and Privacy Shield by the <a href="https://www.cnil.fr/fr/invalidation-du-privacy-shield-les-suites-de-larret-de-la-cjue">Shrems rulings</a>, the controversies surrounding the hosting of the French <a href="https://www.silicon.fr/health-data-hub-pousse-communiquer-benchmark-470295.html">Health Data Hub</a> and unicorn <a href="https://www.latribune.fr/technos-medias/innovation-et-start-up/comment-doctolib-justifie-le-choix-du-geant-americain-amazon-pour-heberger-les-donnees-des-francais-919359.html">Doctolib</a>, the <a href="https://www.ncsc.nl/documenten/publicaties/2022/augustus/16/cloud-act-memo">real impact of US extraterritorial laws</a>, the <a href="https://www.numerique.gouv.fr/services/cloud/doctrine/">French government’s “Doctrine Cloud”</a> and its “Trusted Cloud” taxonomy, vulnerability to <a href="https://www.businessinsider.com/google-locks-afghan-government-email-accounts-to-block-taliban-report-2021-9">US suppliers</a>, etc.</p><p>States and companies are today questioning their dependencies and the trust they can place in their partners, suppliers and the supply chains that link them. The digital sector is no exception to these questions, and is also living through geopolitical upheavals and the quest — rightly or wrongly — for sovereignty.</p><h3>Case study from Amazon</h3><p>How do Cloud providers react to this situation? All Cloud providers are playing a game of transparency and commitment to the confidentiality of their customers’ data, in order to gain and maintain their customers’ trust.</p><p>For AWS, this profession of faith takes the form of half-yearly statistics on the various law enforcement information requests processed by the company over the last 6 months.</p><p>This little-publicized, little-discussed exercise has been carried out for 8 years now: the first report was published for the first half of 2015.</p><p>The data, provided in PDF format and buried in the deep tree structure of amazon.com (you’ll find it <a href="https://www.amazon.com/gp/help/customer/display.html?ref_=help_search_1-2&amp;nodeId=GYSDRGWQ2C2CRYEF&amp;qid=1678961283369&amp;sr=1-2">here</a>) opens up to anyone who wants the possibility of analyzing and measuring (within the limits of what is provided) over time the interactions of AWS and Amazon with US and foreign authorities on this sensitive subject.</p><p>The format and breakdown of the data communicated changed completely in the first half of 2020, forcing a different analysis between its eras.</p><p><strong>Let’s take a closer look!</strong></p><h4>Before 2020</h4><p>From 2015 to the first half of 2020, Amazon provides details by type of judicial requisition, including:</p><ul><li>“<em>Subpoenas</em>”: injunctions issued by various U.S. authorities (courts, attorneys, agencies, juries) without prior substantive review by a judge or magistrate.</li><li><em>“Search warrants”</em>: search warrants issued by courts (local, state, federal) specifically identifying the location of the search and the objects/data to be seized.</li><li><em>“Other court orders”</em>: all other court orders that are neither Subpoenas nor Search Warrants.</li><li><em>“National security requests”</em>: the subject of all fantasies and questions, these are requisitions that fall under the NSL (“National Security Request”) and, above all, the notorious and little-known FISA (“Foreign Intelligence Surveillance Act”) mentioned in the introduction.</li><li><em>“Non-US requests”</em>: all requests from outside the United States.</li></ul><p>While Amazon explains that it only delivers metadata (“<em>Non-Content</em>” in the reporting taxonomy) in the case of Subpoenas, all other requisitions may be subject to delivery of “real” content (“<em>Content</em>”) and/or metadata (“<em>Non-Content</em>”). Until 2020, Amazon does not provide details on the split between “<em>Content</em>” and “<em>Non-Content</em>” responses.<br>Over this 5-year period, however, Amazon differentiates according to the amount of data actually provided to authorities in response to requisitions (this distinction will disappear in reports after 2018). Amazon distinguishes three options:</p><ul><li><em>“Full response”</em>: all requested information is provided</li><li><em>“Partial Response”</em>: only part of the requested information is returned</li><li><em>“No Response”</em>: no information is provided by Amazon.</li></ul><p><a href="https://lookerstudio.google.com/embed/reporting/320d398c-9a8c-4de9-80cf-9d390912b033/page/tEnnC">Amazon &quot;Law Enforcement Information Requests&quot; history</a></p><p><em>📊 All data for this period can be found in the Looker Studio dashboard.</em></p><p>A few lessons learned over the period:</p><ul><li>Total requisition demand increased almost <strong>fourfold</strong>, from 983 in the first half of 2015 to 3,644 in 2020.</li><li><em>Subpoenas</em> are by far <strong>the most represented</strong>: around 18,000 over the period against 24313 requests in total.</li><li>As a reminder, Amazon explains that it only provides metadata for <em>Subpoenas</em> (For <a href="https://www.zulma.fr/livre/lage-du-capitalisme-de-surveillance/">Shoshana Zuboff</a> readers, you know that metadata alone is already a rich mine of information).</li><li>While <strong>Amazon tends to respond quite favorably to requests from American institutions</strong> (only around 25% of requests are met with a “<em>No Response</em>”), the same cannot be said for foreign requests, which are met with 86% “<em>No Response</em>”.</li></ul><h4>After 2020</h4><p>From 2020 onwards, Amazon is changing its reporting:</p><ul><li>The type of court requisition is no longer indicated: “<em>Subpoenas</em>”, “<em>Search Warrants</em>” and “<em>Other court orders</em>” are aggregated in the same indicator.</li><li>A distinction by country of origin of the requisition is added: this is the most interesting part!</li><li>The level of response has been removed (<em>Full</em> / <em>Partial</em> / <em>No response</em>).</li><li>The proportion of data type provided in response is specified: “<em>Content</em>” or “<em>Non Content</em>”.</li></ul><p><a href="https://lookerstudio.google.com/u/0/reporting/320d398c-9a8c-4de9-80cf-9d390912b033/page/p_anqrj5ob4c/edit">Amazon &quot;Law Enforcement Information Requests&quot; history</a></p><p><em>📊 All 2020+ data are in the second tab of the Looker Studio dashboard</em></p><p>Here again, a few lessons to be learned:</p><ul><li>Historically, <strong>the US has been fond of data hosted by AWS</strong> (up to 70% of requests in 2021S1), and Germany fond of data hosted by Amazon (45% of the total).<br>However, a sharp increase in data requisitions from a growing number of countries tends to contradict these proportions (the emergence of India, for example).</li><li>While data requisitions are stable for Amazon, they are increasing for AWS: <strong>x2 in 2 years</strong>.</li><li>Even so, <strong>the vast majority of data requested concerns Amazon</strong> (97.3%), not AWS.</li></ul><p>Another peculiarity is that, at the very end of the document, Amazon now answers this question:</p><blockquote>« How many requests resulted in the disclosure to the U.S. government of enterprise content data located outside the United States? »</blockquote><p>This information is regularly highlighted by AWS in response to customer concerns (for example, here at the Paris <a href="https://youtu.be/MXAy1aaPkqo?si=pdwxhSI93c4Emfzm">AWS Summit 2023</a>). The answer to this question is the same in every half-yearly report: “<strong><em>None</em></strong>”.</p><h4>About National Security Requests</h4><p>Since the first report in 2015 and until today, Amazon cannot legally indicate the exact number of NSL or FISA requests received. Only a range is indicated, and for each half-year since publications began, <strong>this range has been between 0 and 249 requests</strong>.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*4w73sUdlrnL-kMYW7R5oaQ.png" /></figure><h3>Conclusions</h3><p>Despite this abundance of data, it is very difficult to draw conclusions from these reports. Take, for example, the transition between the two periods:</p><ul><li>In the first half of 2020, <strong>3,644 requisitions are reported</strong>, including 5% classified as “Non US requests” (i.e. 195).</li><li>In the second half of 2020, <strong>28,187 requisitions are reported</strong>, 88% of them outside the United States (i.e. 24,737, almost half of them from Germany)!</li></ul><p>Amazon provides no explanation for this sudden inflation, whose main cause is undoubtedly its counting method.</p><p>Taken at face value, these figures tend to minimize the risks of “legal misappropriation” of data stored at AWS by US administrations: little data provided (and when it is, mainly “<em>Non content</em>”), few requisitions concerning AWS, requests mainly from foreign countries…</p><p>However, a number of grey areas remain, particularly around NSLs (including FISA-based requests) and the absence of any specific mention of the CLOUD Act. Let’s not forget that the FISA Act seems to be a fine-meshed net with a rather vast scope, it is also accompanied by <a href="https://www.cairn.info/revue-securite-et-strategie-2013-3-page-64.htm">a culture of secrecy in its extraterritorial application</a> that makes any exercise in transparency inherently tricky, for example:</p><ul><li>Injunctions may be accompanied by a ban on notifying data holders</li><li>European data holders cannot benefit from any of the American constitutional guarantees of protection.</li><li>Individuals who reveal the existence of data collection and surveillance systems may be subject to prosecution and potentially heavy penalties.</li></ul><p>With its “<a href="https://press.aboutamazon.com/2023/10/amazon-web-services-to-launch-aws-european-sovereign-cloud">AWS European Sovereign Cloud</a>” announced for October 2023, AWS is once again trying to show its white coat to Europe and convince the undecided and the perplexed. It’s up to the customer to protect his or her most sensitive data, possibly by choosing another partner, but more pragmatically and systematically by encrypting it, whatever the provider!</p><h3>🛠️ Open data</h3><p>All “<em>raw</em>” data from AWS reports since 2015 is freely available.</p><p>You can find all the information here :</p><ul><li>A <a href="https://console.cloud.google.com/bigquery?project=amazonlawinformationrequest&amp;supportedpurview=project&amp;ws=!1m5!1m4!4m3!1samazonlawinformationrequest!2sFullLawInformationRequest!3s1-LawEnforcementRequest-TS">public dataset</a> available on Google BigQuery</li><li>A <a href="https://lookerstudio.google.com/reporting/320d398c-9a8c-4de9-80cf-9d390912b033">public Looker Studio</a> dashboard available (based on the previous dataset)</li><li>A <a href="https://gitlab.com/casaubon/amazon-law-information-request">GitLab project</a> including raw data in CSV format (and containing the CI/CD feeding the dataset)</li></ul><p><a href="https://gitlab.com/casaubon/amazon-law-information-request">Antoine Lagier / Amazon Law Information Request · GitLab</a></p><p>For more information on tooling, see the <a href="https://gitlab.com/casaubon/amazon-law-information-request/-/blob/main/README.md?ref_type=heads">README in the GitLab project</a>.</p><p>A french version of this paper is also available <a href="https://timspirit.fr/articles/historique-des-requisitions-judiciaires-de-donnees-chez-amazon-et-aws-law-enforcement-information-requests/">here</a>.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=c0fcb478117b" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Overview of Environmental Indicators in the Cloud]]></title>
            <link>https://medium.com/@a.lagier/overview-of-environmental-indicators-in-the-cloud-48c951976721?source=rss-65b5dacc42a6------2</link>
            <guid isPermaLink="false">https://medium.com/p/48c951976721</guid>
            <category><![CDATA[sustainability]]></category>
            <category><![CDATA[gafam]]></category>
            <category><![CDATA[environmen]]></category>
            <category><![CDATA[cloud]]></category>
            <dc:creator><![CDATA[Antoine Lagier]]></dc:creator>
            <pubDate>Tue, 05 Sep 2023 06:59:10 GMT</pubDate>
            <atom:updated>2023-09-06T11:10:18.107Z</atom:updated>
            <content:encoded><![CDATA[<p>Reports follow one another, and the observations remain the same: In march 2023, <a href="https://en.arcep.fr/uploads/tx_gspublication/press-kit-study-Ademe-Arcep-lot3_march2023.pdf">ADEME and ARCEP</a>, in May, the <a href="https://theshiftproject.org/wp-content/uploads/2023/05/Note-danalyse_Planifier-la-decarbonation-du-systeme-numerique-en-France_Mai2023.pdf">Shift Project</a> , the environmental impact of digital technology is spiraling out of control.</p><p>The “business as usual” scenario estimated by ADEME in its 2023 prospective study predicts the carbon footprint of digital technology in France to reach 49 MtCO2eq in 2050, compared to 17 MtCO2eq in 2020 (equivalent to 2.5% of national emissions at that date).</p><blockquote>The digital sector’s share would account for over 60% of the total national emissions by 2050, assuming the country manages to achieve “carbon neutrality” (with an emission quota of around 80 MtCO2eq)!</blockquote><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*OTiaWsomAgmqfoYGSV8Rhg.png" /></figure><p>Faced with this observation, how can we take action? Let’s consider the specific case of your cloud provider. Can they be a partner in your Sustainable Digital approach? What can they offer you?</p><h3>Datacenter Environmental Indicators</h3><p>The datacenter industry has long offered an energy efficiency indicator called PUE (Power Usage Effectiveness). More recently, water consumption indicators (WUE) and carbon intensity indicators (CUE) have also gained attention.</p><p>Let’s take a look at these indicators for the GAFAMs and a few French cloud players.</p><h3>PUE | Power Usage Effectiveness</h3><p>While PUE is a familiar term, there are ongoing discussions about its calculation method and its usefulness in comparing datacenters and Cloud providers.</p><p>PUE measures the energy efficiency of a data center by comparing the total energy consumption of the building to the energy consumption of the equipment (it has no unit). The closer it is to 1, the better, indicating that most energy is directed towards active equipment.</p><p>However, PUE does not consider the carbon intensity of the electricity used, equipment utilisation rates, location impact, etc. Therefore, it cannot be the sole decision criterion and provides only a partial view of data center energy efficiency.</p><p>Furthermore, the <a href="https://www.iso.org/standard/63451.html">PUE standard</a> comes in 4 versions, from PUE0 to PUE3, based on granularity and methodology. It’s often challenging to determine which PUE providers are referring to, especially American ones. These methodological biases are highlighted in this <a href="https://www.podcastics.com/podcast/episode/cloud-souverain-ecologiquement-responsable-136741/">podcast</a> by Yann Lechelle, former CEO of Scaleway, who underlines the differences in calculations between American and European players.</p><h4>State of the Market</h4><p>The evolution of PUE over time is well documented, notably by the <a href="https://journal.uptimeinstitute.com/is-pue-actually-going-up/">Uptime Institute</a>.</p><p>PUE has significantly decreased over the years. (from 2.5 in 2007 to <a href="https://journal.uptimeinstitute.com/is-pue-actually-going-up/">1.67 in 2019</a> and <a href="https://uptimeinstitute.com/about-ui/press-releases/uptime-institute-11th-annual-global-data-center-survey">1.58 in 2021</a>), which has helped keep data center energy consumption growth to <a href="https://science.sciencemag.org/content/367/6481/984.full">just 6% between 2010 and 2018</a>.</p><h4>How Providers Compare</h4><p>Except for Scaleway and Oracle, PUE values are usually communicated only in aggregated and annual forms, hiding potential disparities in time and space (location, host, seasons).</p><p>All providers fall within similar ranges, between 1.10 and 1.30, well below the benchmarks for traditional private ddatacenters as reported by the Uptime Institute.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*OfNotQPB3WL9L7XZ350_BQ.png" /><figcaption>Overview of IT players’ PUE</figcaption></figure><p>Scaleway has made a further effort to improve transparency, publishing a complete dashboard on the behavior of each datacenter, including real-time PUE (the two PUE values for Scaleway concern older datacenters -before 2018- and newer ones -after 2018-) :</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/744/0*asgP2Ec5zmOD-i4S.png" /><figcaption>Scaleway DC3 dashboard (<a href="https://pue.dc3.scaleway.com/">https://pue.dc3.scaleway.com/</a>)</figcaption></figure><h3>WUE | Water Usage Effectiveness</h3><p>Introduced in the early 2010s, WUE highlights the amount of water used relative to the electricity consumed by equipment. Its unit is l/kWh (liters per kilowatt-hour), and lower values are better.</p><p>While PUE and energy are at the top of the list of indicators observed for Datacenters, recent months have seen a rise in concerns about the sometimes gargantuan water consumption of Datacenters, notably by <a href="https://www.theregister.com/2022/12/19/google_datacenters_dalles/">Google in The Dalles</a>, and in general about the <a href="https://www.cnbc.com/2022/11/15/microsoft-meta-others-face-rising-drought-risk-to-their-data-centers.html">risks associated with drought</a>.</p><p>The 3 American hyperscalers have announced a whole range of “water-positive” commitments: by 2030, they will be giving back as much water (or more, in the case of Google!) as they use for their operations.</p><h4>State of the Market</h4><p>Since WUE is not as well known and adopted as PUE, there are no similar historical data. Nevertheless, according to the US Department of Energy (DoE), the average US datacenter consumes <a href="https://www.datacenterknowledge.com/archives/2016/07/12/heres-how-much-water-all-us-data-centers-consume">around 1.8 liters of water per kWh</a>.</p><h4>How Providers Compare</h4><p>All the French and American players are positioned well below the market standard with the exception of Microsoft, which tops out at 0.49 l/kWh, and Google, for which no figures seem to be available.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Yo9ljfVo6dJ9-jy0U1NQQw.png" /><figcaption>Overview of IT players’ WUE</figcaption></figure><h3>CUE | Carbon Usage Effectiveness</h3><p>The last indicator to be reviewed, after energy and water, is carbon of course!</p><p><a href="https://www.netalis.fr/wp-content/uploads/2016/04/Carbon-Usage-Effectiveness-White-Paper_v3.pdf">The CUE</a> relates the quantity of CO2 emitted per kWh of energy supplied to the equipment, so its unit is CO2eq/kWh. Like the WUE, both of which were formalized together in the early 2010s, the more it tends towards 0, the more it means that the amount of carbon emitted will be optimized.</p><p>Paradoxically, at a time when press headlines are bursting with the injunction to reduce emissions, this indicator is conspicuously absent from all cloud publisher publications, with the notable exception of OVH.</p><p>In its place comes an avalanche of commitments of all kinds: “carbon neutral”, “100% energy renewable”, “carbon free energy 24/7”, “net zero carbon” and “carbon negative”!</p><p>It’s hard to know how serious these commitments are, given that carbon accounting authorizes a wide variety of different methods (market-baset/location-based, GHG Protocol, BEGES, etc...), with widely varying perimeters taken into account (the famous scope 3). Others have already carried out this analysis, and I strongly recommend that you read it (for example, here at <a href="https://www.carbone4.com/en/analysis-carbon-footprint-cloud">Carbone 4</a>, with the <a href="https://youtu.be/tmeiJQbL9Xo">webinar version available here in video</a>, or at our friends from <a href="https://www.boavizta.org/en">Boavizta</a>).</p><p>What’s more, these commitments do little to hide a vague definition of what neutrality or offsetting can be, particularly at company level. On this subject, read <a href="https://librairie.ademe.fr/developpement-durable/5335-utilisation-de-l-argument-de-neutralite-carbone-dans-les-communications.html">Ademe’s July 2021 note</a> on the scientific and methodological reality of carbon neutrality: it can only apply to a state or a planet, not a company. Similarly, recent studies tend to show the inanity of offset offers currently on the market (see, for example, these Guardian articles <a href="https://www.theguardian.com/environment/2023/jan/18/revealed-forest-carbon-offsets-biggest-provider-worthless-verra-aoe">here</a> and <a href="https://www.theguardian.com/environment/2023/may/24/chevron-carbon-offset-climate-crisis">here</a>).</p><h4>State of the Market</h4><p>Unfortunately, there’s no such thing as an average or typical CUE in the literature (unless I’m mistaken!).</p><h4>How Providers Compare</h4><p>Only OVH publicly displays a CUE of 0.20 tCO2eq/MWh.</p><p>No other figures for any other player are available. In view of this scarcity of measurements, let’s try to summarize the various commitments made by suppliers:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*X_D5ELOVveEReYQ_Nk7IPg.png" /><figcaption>Overview of IT players’ carbon commitments</figcaption></figure><p>If we summarise these commitments, we can see that the various players share a number of common points:</p><ul><li>“<em>Carbon neutral</em>” : This point of passage seems to agree on a definition where emissions are offset by compensation projects (avoidance and/or capture). For <a href="https://blogs.microsoft.com/blog/2020/01/16/microsoft-will-be-carbon-negative-by-2030/">Microsoft</a>, this means “<em>offsetting their emissions with payments either to avoid a reduction in emissions or remove carbon from the atmosphere</em>”.<br>On this first point of passage, we note a few absentees, such as AWS.</li><li>“<em>100% renewable</em>” : Contrary to what you might think, this is not a commitment to consume 100% renewable electricity, whatever the time of day (see Google’s peculiarity on this subject below).<br>Rather, it’s a commitment that for every kWh consumed, 1 kWh of green electricity will be produced, but probably <em>elsewhere</em> on the planet.<br>Consequently, <strong>you cannot deduct from this commitment the actual carbon intensity consumed by your services</strong> during their use, as it is in no way representative of the carbon intensity of the electrical network to which the datacenter is connected.</li><li>“<em>Net-zero carbon</em>” : this commitment is clearly linked to the objectives of the Paris Agreement and of global carbon neutrality by 2050. In this case, no offsets purchased on the market, but carbon emissions naturally absorbed by the planet’s natural capacities.<br>It is clearly in this commitment that the contradiction highlighted by Ademe in its note is most evident.</li></ul><p>As confusion reigns over the methodological definition of these objectives, it is not impossible (or even probable?) that each player has its own definition of what it means by these various terms. However, as these definitions are never clearly spelled out, these commitments are obviously to be taken with a great deal of mistrust and caution.</p><p>In addition, some providers are displaying particularities that are rarely found among the others:</p><ul><li><strong>Google</strong> and <strong>Microsoft</strong> have committed to a real, 24x7 consumption of green electricity wherever they operate by 2030. This is an improved, “real-time” version of the previous “100% renewable” commitment.<br>Both Google and Microsoft are signatories to the <a href="https://gocarbonfree247.com/our-partners/">Go Carbon Free 24x7 initiative</a>.</li><li><strong>Microsoft</strong>, for its part, has the most ambitious objective, since they consider that by 2050 they will have removed from the atmosphere all the carbon emitted by the company since its creation…<br>“<strong>Promises only bind those who believe them” !</strong></li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*1KmQ0uVr64MpDQtYqq6AXw.png" /></figure><h3>Conclusions</h3><p>The exercise of collecting and comparing environmental indicators and commitments from different Cloud providers is made tricky by the vagueness of the methodology and the grandiose claims made by certain players.</p><p>Clearly, these industrial players are demonstrating a certain efficiency in the management of their Datacenters and services, with PUE and WUE levels achieved that are without exception far better than those of ordinary private Datacenters.</p><p>We are convinced that <strong>a reasonable and informed use of Public Cloud</strong> services is an important ally from an environmental point of view, but that the requirements necessary for this “reasonable” use are unfortunately not yet in place.</p><p>Using one of these providers in no way relieves you of your responsibility to make a “fair” use of the resources you deploy.</p><blockquote>The Cloud, beneath its alluring trappings and ethereal elasticity, is based on a very real materiality.</blockquote><p>Even before deploying your first resource, you can make structuring choices, notably in the choice of region (<a href="https://app.electricitymaps.com/map">carbon intensity in France can be more than 15 to 20 times lower than in Ireland</a>), before using your supplier’s resources without falling into the usual pitfalls (over-quality, lack of measurement or compulsive hoarding!).</p><p>Similarly, it’s up to us, as members of the ecosystem gravitating around these Cloud Providers, to lobby them more than ever to obtain greater transparency and leverage from them.</p><h3>References</h3><h4>AWS</h4><ul><li><a href="https://aws.amazon.com/blogs/industries/four-trends-driving-global-utility-digitization/">https://aws.amazon.com/blogs/industries/four-trends-driving-global-utility-digitization/</a></li><li><a href="https://sustainability.aboutamazon.com/aws-water-positive-methodology.pdf">https://sustainability.aboutamazon.com/aws-water-positive-methodology.pdf</a></li><li><a href="https://press.aboutamazon.com/2022/11/aws-makes-water-positive-commitment-to-return-more-water-to-communities-than-it-uses-by-2030">https://press.aboutamazon.com/2022/11/aws-makes-water-positive-commitment-to-return-more-water-to-communities-than-it-uses-by-2030</a></li><li><a href="https://sustainability.aboutamazon.com/environment/the-cloud?energyType=true">https://sustainability.aboutamazon.com/environment/the-cloud?energyType=true</a></li></ul><h4>AZURE</h4><ul><li><a href="https://azure.microsoft.com/en-us/blog/how-microsoft-measures-datacenter-water-and-energy-use-to-improve-azure-cloud-sustainability/">https://azure.microsoft.com/en-us/blog/how-microsoft-measures-datacenter-water-and-energy-use-to-improve-azure-cloud-sustainability/</a></li><li><a href="https://azure.microsoft.com/en-us/blog/how-microsoft-measures-datacenter-water-and-energy-use-to-improve-azure-cloud-sustainability/">https://azure.microsoft.com/en-us/blog/how-microsoft-measures-datacenter-water-and-energy-use-to-improve-azure-cloud-sustainability/</a></li><li><a href="https://blogs.microsoft.com/blog/2020/09/21/microsoft-will-replenish-more-water-than-it-consumes-by-2030/">https://blogs.microsoft.com/blog/2020/09/21/microsoft-will-replenish-more-water-than-it-consumes-by-2030/</a></li><li><a href="https://blogs.microsoft.com/blog/2020/01/16/microsoft-will-be-carbon-negative-by-2030/">https://blogs.microsoft.com/blog/2020/01/16/microsoft-will-be-carbon-negative-by-2030/</a></li><li><a href="https://blogs.microsoft.com/blog/2021/07/14/made-to-measure-sustainability-commitment-progress-and-updates/">https://blogs.microsoft.com/blog/2021/07/14/made-to-measure-sustainability-commitment-progress-and-updates/</a></li></ul><h4>GCP</h4><ul><li><a href="https://www.gstatic.com/gumdrop/sustainability/google-2022-environmental-report.pdf">https://www.gstatic.com/gumdrop/sustainability/google-2022-environmental-report.pdf</a></li><li><a href="https://sustainability.google/commitments/water/#water-at-google">https://sustainability.google/commitments/water/#water-at-google</a></li><li><a href="https://www.blog.google/outreach-initiatives/sustainability/carbon-neutrality-by-end-of-2007/">https://www.blog.google/outreach-initiatives/sustainability/carbon-neutrality-by-end-of-2007/</a></li><li><a href="https://www.blog.google/outreach-initiatives/environment/meeting-our-match-buying-100-percent-renewable-energy/">https://www.blog.google/outreach-initiatives/environment/meeting-our-match-buying-100-percent-renewable-energy/</a></li><li><a href="https://sustainability.google/progress/energy/">https://sustainability.google/progress/energy/</a></li></ul><h4>META</h4><ul><li><a href="https://sustainability.fb.com/wp-content/uploads/2022/06/Meta-2021-Sustainability-Report.pdf">https://sustainability.fb.com/wp-content/uploads/2022/06/Meta-2021-Sustainability-Report.pdf</a></li><li><a href="https://about.fb.com/news/2021/08/restoring-water">https://about.fb.com/news/2021/08/restoring-water</a></li><li><a href="https://tech.facebook.com/engineering/2021/4/renewable-energy/">https://tech.facebook.com/engineering/2021/4/renewable-energy/</a></li><li><a href="https://sustainability.fb.com/climate/">https://sustainability.fb.com/climate/</a></li></ul><h4>OVH</h4><ul><li><a href="https://corporate.ovhcloud.com/fr/sustainability/environment/">https://corporate.ovhcloud.com/fr/sustainability/environment/</a></li><li><a href="https://www.computerweekly.com/news/252491498/OVHcloud-doubles-down-on-sustainability-with-carbon-reduction-pledges">https://www.computerweekly.com/news/252491498/OVHcloud-doubles-down-on-sustainability-with-carbon-reduction-pledges</a></li></ul><h4>SCALEWAY</h4><ul><li><a href="https://www.scaleway.com/en/environmental-leadership/">https://www.scaleway.com/en/environmental-leadership/</a></li></ul><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=48c951976721" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[2011–2021 A decade of public cloud]]></title>
            <link>https://medium.com/@a.lagier/2011-2021-a-decade-of-public-cloud-b6cd58a5e6b2?source=rss-65b5dacc42a6------2</link>
            <guid isPermaLink="false">https://medium.com/p/b6cd58a5e6b2</guid>
            <category><![CDATA[iaas]]></category>
            <category><![CDATA[cloud]]></category>
            <dc:creator><![CDATA[Antoine Lagier]]></dc:creator>
            <pubDate>Sat, 27 Nov 2021 14:08:29 GMT</pubDate>
            <atom:updated>2026-11-03T21:40:45.602Z</atom:updated>
            <content:encoded><![CDATA[<h3>2011–2021 (and beyond) : A decade of public cloud</h3><p>An animated gif history</p><ul><li><strong>2023 MQ gif </strong><a href="https://www.linkedin.com/posts/alagier_cloud-iaas-activity-7162766156880166912-q7L9?utm_source=share&amp;utm_medium=member_desktop"><strong>here on LinkedIn</strong></a></li><li><strong>2024 MQ gif </strong><a href="https://www.linkedin.com/posts/alagier_aws-gcp-azure-activity-7287009458675675136--R3a?utm_source=share&amp;utm_medium=member_desktop"><strong>here on Linkedin</strong></a></li><li><strong>2026 MQ gif </strong><a href="https://lnkd.in/p/eQuAPgdP"><strong>here on Linkedin</strong></a></li></ul><p>The 2010s are over!</p><p>Gartner recently published its “<a href="https://www.gartner.com/technology/media-products/reprints/AWS/1-271W1OSK-FRA.html">Magic Quadrant for Cloud Infrastructure and Platform Services</a>” for the year 2021, bringing the 2010–2020 period to a close and marking the entry into the new 2021/2030 decade.</p><p>An opportunity to look back and measure the progress of the Cloud industry against these publications.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*PyDfZEVMDfCM8lKPFNDDfA.gif" /><figcaption><em>10 years of Public Cloud in 1 animated gif !</em></figcaption></figure><h3>Back to 2010</h3><p>Who remembers the ancestor of this Quadrant, which we wait and browse every year as we wait for the result of the trifecta? In 2010, it was called “<a href="http://www.mindyourservice.com/assets/docs/Gartner_Magic_Quadrant_2010.pdf">Magic Quadrant for Cloud Infrastructure as a Service and Web Hosting </a>” and included no less than 20 players!</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/350/1*dtNaQI77LbAsIDNzoP8wAw.jpeg" /><figcaption><em>Who remembers the Quadrant from 10 years ago?</em></figcaption></figure><p>The five leaders in the quadrant at the time were Savvis (later bought by CenturyLink), AT&amp;T, Rackspace, Verizon, and Terremark (bought by Verizon the following year). They have all quickly disappeared from the radar since then, and have even left the IaaS market altogether.</p><p>Nevertheless, some of them are familiar: SoftLayer and IBM are already there, but it is above all AWS (still “Amazon” on the Quadrant!) that is already showing itself to be the most visionary (but far from being the most capable of delivering), and this will be the only year in which it will not be at the very top right of the Quadrant.</p><p>Gartner’s analysis of Amazon at that time is worth a look, judge for yourself:</p><blockquote>“[Amazon] is a thought leader; it is extraordinarily innovative, exceptionally agile, and very responsive to the market. […] Amazon is a price leader, but it charges separately for optional items that are often bundled with competitive offerings. <strong>Prospective customers should be careful to model the costs accurately, especially network-related charges</strong>.”<br>Magic Quadrant for Cloud Infrastructure as a Service and Web Hosting, Gartner, December 2010</blockquote><p>Still relevant 10 years later! It should be noted that <a href="https://medium.com/timspirit/whats-next-for-finops-2800a2286c9d">the FinOps discipline</a>, still unknown at the time, was already in the making and had a bright future.</p><p>The other two major hyperscalers we know today arrived later. Unlike AWS, these two first focused on the public cloud through PaaS:</p><ul><li><a href="https://azure.microsoft.com/en-in/blog/microsoft-announces-windows-azure-and-azure-services-platform/">Windows Azure was announced in 2008</a> before being GA in 2010 and positioning itself on IaaS from 2012 with its <a href="https://www.techrepublic.com/blog/the-enterprise-cloud/microsoft-azure-now-offers-iaas-to-compete-with-rackspace-amazon-in-public-cloud-space/">first virtual machine services</a></li><li>Similar development for GCP, which <a href="http://googleappengine.blogspot.com/2008/04/introducing-google-app-engine-our-new.html">announced Google App Engine in 2008</a> before positioning itself in IaaS a few years later at the end of 2013 with Compute Engine.</li></ul><p>Both will arrive in the Quadrant at the bottom right before joining AWS at the top: quickly in 2013 for Microsoft and only in 2018 for GCP. This late entry of Google among the leaders also illustrates the difficulty that Google has in establishing itself on the market as a credible player.</p><p>For example, in 2019, the report states:</p><blockquote>“Google demonstrates an immaturity of process and procedures when dealing with enterprise accounts, which can make the company difficult to transact with at times”<br>Magic Quadrant for Cloud Infrastructure as a Service, Worldwide, Gartner, December 2019</blockquote><p>A subject still relevant in 2021 when <a href="https://www.usinenouvelle.com/article/google-va-t-il-etre-contraint-de-quitter-le-marche-du-cloud-en-2023.N1155687?">its long-term presence in the industry</a> continues to raise questions.</p><h3>2012–2017: the emergence of a market</h3><p>Nothing and no one can escape it, the Public Cloud is making its mark in all sectors, on all continents, and for all service models (IaaS, SaaS, PaaS). <a href="http://nothing%20and%20no%20one%20can%20escape%20it%2C%20the%20public%20cloud%20is%20making%20its%20mark%20in%20all%20sectors%2C%20on%20all%20continents%2C%20and%20for%20all%20service%20models%20%28iaas%2C%20saas%2C%20paas%29.%20gartner%20estimates%20that%20the%20public%20cloud%20market%20will%20grow%20by%2023%25%20between%202020%20and%202021%2C%20reaching%20332%20billion%20dollars%20this%20year%2C%2082%20billion%20of%20which%20will%20be%20for%20iaas%20alone./">Gartner estimates that the public cloud market will grow by 23% between 2020 and 2021</a>, reaching 332 billion dollars this year, 82 billion of which will be for IaaS alone.</p><p>The emergence of the market can be seen at the beginning of the decade in the first quadrants, with many players and candidates launching themselves into the market that AWS has cleared. In the four years between 2012 and 2015, no less than 24 different names will be analyzed by Gartner, while Alibaba, Oracle, and Tencent, present in the last edition, have not yet joined the arena.</p><p>Many of these early winners will throw in the towel and switched to other models, such as OVH, VMware, and Rackspace. Others will be bought out: Terremark and Softlayer will end up with IBM, CSC with HP.</p><p>Ironically, Rackspace, a historical player in private and public clouds through its significant contribution to OpenStack as a founding member and major maintainer, will even end up focusing on the stacks of its former competitors Azure and VMware in the early 2020s.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/525/1*BLw_awkj4_L8tjk8grgdAw.png" /><figcaption><em>Evolution of Rackspace over the years (watch out for the direction!)</em></figcaption></figure><p>It was also in the mid-2010s that the revenues of the Cloud divisions began to be shown separately in quarterly financial publications, putting an end to years of suspense about the presumed profitability of this industry against a backdrop of price wars.</p><p>Amazon started 2015 with flattering figures that would continue to be verified year after year: AWS was largely profitable (<a href="https://venturebeat.com/2013/09/05/amazons-mountain-of-margin-in-cloud-services-over-80-profit/">with very comfortable margins</a>) and would continue to contribute to the group’s future good results, regularly accounting in recent years for <a href="https://www.techradar.com/news/aws-is-now-a-bigger-part-of-amazons-overall-success-than-ever">more than half of Amazon’s profits</a> for only about 15% of its revenues.</p><h3>2017–2018: refocusing</h3><p>A refocusing will take place in 2017/2018 in two ways:</p><ul><li>Gartner is pushing AWS and Microsoft back in its Quadrant, as they have risen too high, too fast, leaving little room for possible progression.</li><li>Alibaba and Oracle, both still present today, join the Quadrant at this time. 2017 will be the last Quadrant in which all the challengers will be present: the following year, 8 competitors will disappear, i.e. more than half of the companies analysed by Gartner the previous year.</li></ul><p>The leaders are then far away and the barrier to entry is high. The market is about to complete its consolidation: “Winner takes it all” (<a href="https://www.statista.com/chart/7994/cloud-market-share/">or almost</a>).</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/525/1*JiwY6_HRhAs-2v-DNf7REQ.png" /><figcaption>AWS from 2012 to 2021</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/525/1*3xhdR8ybxVbMv0LZlFvv_w.png" /><figcaption>Azure from 2013 to 2021</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/525/1*zrhG4_FGd4gvEWWcKXWHdg.png" /><figcaption>GCP from 2014 to 2021</figcaption></figure><h3>After 2018: maturity?</h3><p>At the midpoint of the decade, the trifecta that we have seen for the past five years is already in place. In 2018, Google, AWS, and Microsoft will all find themselves in the top right-hand quarter of the leaders for the first time.</p><p>Since then, the pack has been split in two: in addition to the American hyperscalers comfortably installed in the top right quadrant, the bottom left quadrant of “niche players” brings together the group of pursuers in 2 categories:</p><ul><li>The 20th century glories IBM and Oracle back in the game</li><li>The new Chinese players Alibaba and Tencent (which came last in 2020)</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/525/1*iYge1G10JDcfMD3ilXRfLA.png" /><figcaption>IBM (and Softlayer) from 2012 to 2021</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/525/1*RT-dTmh74HhJfMLE_yIIbQ.png" /><figcaption>Oracle from 2017 to 2021</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/525/1*85OttePuqoNr4HmvHyCNcg.png" /><figcaption>Alibaba Cloud from 2017 to 2021</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/525/1*CD6YcPcFPDUzTvZ7yqJJMQ.png" /><figcaption>Tencent Cloud in 2020 and 2021</figcaption></figure><p>These 7 actors have been following each other for the past four years in a hierarchy that seems complicated to change, especially when you get closer to the top of the ranking.</p><p>Today, AWS and Azure seem to be indisputable as their adoption by companies, for different reasons, remains a fundamental trend. Google <a href="https://www.fiercetelecom.com/platforms/google-cloud-revenue-jumps-45-but-profit-still-elusive">is slow to catch up in terms of profitability</a> even though GCP has a recognized technology brand and stack.</p><p>Here are a few criteria that could eventually upset the established order:</p><ul><li><strong>Reliability</strong> (availability &amp; security)</li></ul><p>Although the breakdowns that emerge in CSPs are often limited in time and impact, security concerns are nevertheless emerging. They concern Azure in particular after a few complicated months in 2021 (successive faults in <a href="https://www.theverge.com/2021/8/27/22644161/microsoft-azure-database-vulnerabilty-chaosdb">Cosmos DB</a>, <a href="https://arstechnica.com/information-technology/2021/09/poc-exploit-released-for-azure-ad-brute-force-bug-heres-what-to-do/">AD</a>, <a href="https://www.theregister.com/2021/09/09/azure_container_flaw/">ACI</a>). <br>Gartner has been alerting its customers to these aspects of Microsoft for three years (2019, 2020, 2021).</p><ul><li><strong>Sustainability</strong></li></ul><p>The poor relation today in terms of services provided to users. Although CSPs regularly make announcements concerning their greenhouse gas reduction or green electricity generation objectives, few offer their customers the ability to monitor their carbon footprint. <br><a href="https://azure.microsoft.com/en-us/blog/microsoft-sustainability-calculator-helps-enterprises-analyze-the-carbon-emissions-of-their-it-infrastructure/">Azure</a> and <a href="https://cloud.google.com/carbon-footprint">Google</a> have recently provided specific services on this subject but the silence from AWS today is deafening.</p><ul><li><strong>Expertise</strong></li></ul><p>CSPs require a significant investment in terms of technology watch and skills. Services are multiplying and expertise in the market is limited and in demand. The choice of a CSP could be influenced by the ability to find the right human resources in the market.</p><ul><li><strong>Money</strong></li></ul><p>Setting up a public cloud is very expensive. In this respect, not all the members of the Quadrant in 2021 are in the same boat: What can the future hold when Oracle, for example, “<a href="https://www.arnnet.com.au/article/692770/how-much-does-aws-microsoft-google-cloud-spend-cloud/?utm_medium=rss&amp;utm_source=tagfeed">invests as much in one year as AWS does in one week in the last quarter of 2020</a>”?</p><ul><li><strong>Sovereignty</strong></li></ul><p>Will European initiatives succeed in making local players emerge (<a href="https://www.google.com/url?sa=t&amp;rct=j&amp;q=&amp;esrc=s&amp;source=web&amp;cd=&amp;cad=rja&amp;uact=8&amp;ved=2ahUKEwjiut3Pnoz0AhUvyoUKHXbwD5cQFnoECAUQAQ&amp;url=https%3A%2F%2Fwww.silicon.fr%2Fcloud-europeen-euclidia-alternative-gaia-x-412268.html&amp;usg=AOvVaw1gPlFkUUJsdXsRRYMgXQqk">Gaia-X</a>, <a href="https://www.google.com/url?sa=t&amp;rct=j&amp;q=&amp;esrc=s&amp;source=web&amp;cd=&amp;cad=rja&amp;uact=8&amp;ved=2ahUKEwjiut3Pnoz0AhUvyoUKHXbwD5cQFnoECAUQAQ&amp;url=https%3A%2F%2Fwww.silicon.fr%2Fcloud-europeen-euclidia-alternative-gaia-x-412268.html&amp;usg=AOvVaw1gPlFkUUJsdXsRRYMgXQqk">Euclidia</a>), and will Chinese players succeed in imposing themselves outside their Asian sphere of influence?</p><p>See you in 10 years to see what the Quadrant of 2030 will look like!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=b6cd58a5e6b2" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Theory of Constraint’s Dice Game Simulator]]></title>
            <link>https://medium.com/@a.lagier/theory-of-constraints-dice-game-simulator-872e45f08e4c?source=rss-65b5dacc42a6------2</link>
            <guid isPermaLink="false">https://medium.com/p/872e45f08e4c</guid>
            <category><![CDATA[flow]]></category>
            <category><![CDATA[management]]></category>
            <category><![CDATA[lean]]></category>
            <category><![CDATA[theory-of-constraints]]></category>
            <category><![CDATA[goldratt]]></category>
            <dc:creator><![CDATA[Antoine Lagier]]></dc:creator>
            <pubDate>Sat, 23 May 2020 20:42:06 GMT</pubDate>
            <atom:updated>2026-10-19T11:19:57.823Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*5QLiJxg0Bluiu3lh" /><figcaption>Photo by <a href="https://unsplash.com/@rihok?utm_source=medium&amp;utm_medium=referral">Riho Kroll</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a></figcaption></figure><p>If you have read <a href="https://en.wikipedia.org/wiki/The_Goal_(novel)">“The Goal”</a> by <a href="https://en.wikipedia.org/wiki/Eliyahu_M._Goldratt">Eliyahu M. Goldratt</a> (i hope you have), you know that in chapters 13 and 14, the main character in the novel Alex goes with his son on a scout hike.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/237/1*9m1ByysOkmlqznZWaRteUg.jpeg" /></figure><p>It then comes to his mind to put into practice through simulation his first thoughts about how running smoothly his factory for which he is accountable (after some discussions with his adviser Jonah).</p><p>Let’s play this game of bowls and matches to enter the world of the <a href="https://en.wikipedia.org/wiki/Theory_of_constraints">theory of constraints</a>.</p><h3>Rules of game</h3><p>The game is actually an experiment in what could be a <em>balanced plant.</em></p><p>The sequence is:</p><ul><li>Each Work Center (WC) in the plant is represented by a bowl.<br>Each bowl belongs to a kid.</li><li>Matches represent the product (or inventory) processed in the plant.<br>A bowl/Work Center takes the matches as inputs and delivers matches as outputs to the next bowl/WC in the production line.</li><li>Each kid rolls a dice to simulate the performance of his bowl/Work Center (i.e. how many matches it is able to transform and give to the next bowl/Work Center)</li></ul><p>Quite simple right?</p><p>But like any production plant, any given Work Center (N) is dependent on the upstream Work Center (N-1) outputs : matches delivered by the “N-1” bowl will become the inventory of the “N” bowl.</p><p>Two exceptions:</p><ul><li>First Work Center takes raw material as inputs (let say there is no raw material shortage).</li><li>Last Work Center gives manufactured finished goods as outputs, ready to be delivered to the customer.</li></ul><h4>Roll the dice &amp; play</h4><p>Each kid will roll the dice one after another:</p><ol><li>The first one places in his bowl the number given by his dice.<br>If he rolls a 5, that means he transformed 5 items from the raw materials and those 5 items are usable for the next bowl.</li><li>The next one in the production line rolls his dice. Please note that he is limited by the outputs of the previous bowl.<br>For instance, if he rolls a 4, he will place 4 matches in his bowl. On the other hand, if he rolls a 6, he will only place 5 matches in his bowl as the previous bawl has only 5 matches available.</li><li>And so on until the last bowl that will finally give the real production of the line.<br>At this stage, you restart from the first bowl for a new production and you keep the inventory in each one.</li></ol><h3>Google Sheet Simulator</h3><p>You will find here a simulator of this game in a Google Sheet:</p><p><a href="https://docs.google.com/spreadsheets/d/1-RLPknqBAEqrXOuFy_GJsGCjWW_h9PK4c9UgnktnYX0/edit?usp=sharing">Theory of constraints</a></p><p>You can refresh it to simulate the rolls or copy (File &gt; Make a copy) the sheet into your account. You can also directly copy the spreadsheet to your Google account by clicking <a href="https://docs.google.com/spreadsheets/d/1-RLPknqBAEqrXOuFy_GJsGCjWW_h9PK4c9UgnktnYX0/copy">on this link</a>.</p><p>By default, the simulation is done with a 6-sided dice. Simulator includes <strong>5 Work Centers</strong> and <strong>10 rounds of rolls</strong> (i.e. each kid will roll the dice 10 times).</p><h3>Expectations</h3><p>Each kid has a 6-sided dice, they all have the same ability to process outputs from the upstream Work Center into new inputs for the downstream Work Center. That’s why it seems to be a <em>balanced plant</em>.</p><p>We should <em>theoretically</em> expect an average flow of 3.5 for the production line and for each Work Center (the midpoint between the lowest 1 and the highest 6).</p><p>We can then define the performance for each roll in relation to the midpoint : if a Work Center rolls a 2, it will have a performance of -1.5; if it rolls a 6, a performance of 2.5 and so on. Let’s call that <strong>Roll Drift</strong>.</p><h3>Guide</h3><p>The sheet simulates 10 rounds of rolls like this one:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*ysRBKl1TODuEXBVPv1ZrUw.jpeg" /><figcaption>A round of rolls : a run through the whole production line</figcaption></figure><ol><li><strong>WC1</strong> rolls a 3 and places 3 matches in its bowl. The drift for this roll is -0.5.</li><li><strong>WC2</strong> rolls a 1. As it has 3 matches available from WC1, it places 1 match in its bowl for WC3 and adds 2 matches to its inventory. Its drift is -2.5.</li><li><strong>WC3</strong> rolls a 4. Unfortunately, it has only 1 match available from its upstream bowl (WC2) and nothing in its inventory so it only places 1 match in its bowl. Drift of -2.5 as well.</li><li>At the end of this round, the final WC5 rolls a 6! But as it has only 1 match from WC4, <strong>the final production is single match only</strong>. Final drift of -2.5 once again.</li></ol><h3>Observations</h3><p>In addition of the detailed rolls for the 10 rounds, you will find 3 charts at the top of the Google Sheet.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*uKg-JlpHIeXyrQ7hRt4YWQ.png" /><figcaption>3 charts to analyse the production</figcaption></figure><h4>Delivered Cumulative Drift (Roll over Roll)</h4><p>This chart shows the evolution of the drift for the final Work Center, i.e. the actual production for the entire production line. Each roll is added to the previous one (<em>cumulative</em>).</p><p>Here, the first two rounds of rolls ended with a single match production (-2.5 drift), leading to a -5 cumulative. Line 0 represents the midpoint.</p><h4>Inventory sum (Roll over Roll)</h4><p>At each roll, a Work Center potentially add some matches to its inventory. This chart shows the evolution of the inventory roll after roll for each Work Center.</p><h4>Cumulative drift per Work Center</h4><p>This shows the cumulative drift for each Work Center roll after roll.</p><h3>Results</h3><p>You can simulate the rolls indefinitely : it is very unlikely to expect an average throughput consistent with the midpoint of the dice (3.5 for a 6-sided dice).</p><p>You will always end up with:</p><ul><li><strong>A never ending drop in cumulative drift</strong>.<br>The production line will never stop delivering less that you might expect.</li><li><strong>An increase in inventory</strong>.<br>Matches stack up in the bowls.</li></ul><h3>Explanations</h3><p>Theory of Constraints aims to find and remove bottleneck(s) in a production line, following the principle that “a chain is no stronger than its weakest link”.</p><p>Even if the plant seems balanced (same maximum capacity for each work center here), you will end up with some bottlenecks somewhere restricting the throughput of the entire system.</p><p>Theory of Constraints focuses on 3 key considerations:</p><ul><li><strong>Throughput</strong><br>The rate at which you can deliver and that generate revenue.<br>“The Goal” is to increase <strong>throughput</strong>.</li><li><strong>Inventory</strong><br>Anything used to produce finished goods : raw materials, facilities, equipment and work in progress (WIP).<br>“The Goal” is to remove <strong>inventory</strong>.</li><li><strong>Operational expenses<br></strong>Money spent by the company to turn inventory to throughput.<br>“The Goal” is to reduce <strong>Operational Expenses</strong>.</li></ul><p>In our case, without managing the constraint (the bottleneck), we have seen that throughput drops and inventory increased.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=872e45f08e4c" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[What’s next for FinOps ?]]></title>
            <link>https://medium.com/@a.lagier/whats-next-for-finops-2800a2286c9d?source=rss-65b5dacc42a6------2</link>
            <guid isPermaLink="false">https://medium.com/p/2800a2286c9d</guid>
            <category><![CDATA[financial-management]]></category>
            <category><![CDATA[finops]]></category>
            <category><![CDATA[cloud]]></category>
            <category><![CDATA[information-technology]]></category>
            <category><![CDATA[cloud-computing]]></category>
            <dc:creator><![CDATA[Antoine Lagier]]></dc:creator>
            <pubDate>Wed, 06 May 2020 08:15:28 GMT</pubDate>
            <atom:updated>2020-05-19T12:12:35.412Z</atom:updated>
            <content:encoded><![CDATA[<h3>What’s next for FinOps ? FinDev !</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*eSV6SRL-QoEgIEs_kE4T1Q.jpeg" /></figure><p>Mindset, tools, skills, best practices… Today, the wide and new playground of financial optimization in the pay-as-you-go domain is being covered from all sides: new players and frameworks, books, training, certifications and so on…</p><p>A <a href="http://finops.org/">FinOps Foundation</a> was even created in the US in February in 2019 (what would IT do without all its foundations?) as well as the first <a href="https://www.meetup.com/Meetup-Cloud-FinOps-Paris/">French Meetup group on the topic</a>, of which <a href="http://timspirit.fr/">Timspirit</a> is glad of being a founding member and which is waiting for you!</p><p>The relevance of cost control is obviously not to be questioned. At a time when large and small organizations are struggling to define their FinOps approach practically, what could we say about its future if we raised our heads ?</p><h3>Go back to the origin first</h3><p>The optimization and control of expenses did not emerge with the Cloud, or even pay-as-you-go. So why have this term and discipline become so popular in recent years?</p><p>In this paper, let’s look at one of the source of FinOps. In this common-sense accounting assertion, let’s look at the first part: <strong>IT resources usage</strong>.</p><figure><img alt="Basic assertion" src="https://cdn-images-1.medium.com/max/1024/1*TEUVqhV21xdqC8SjXEBOKw.png" /><figcaption>Basic FinOps assertion</figcaption></figure><p>We all have this kind of curve in mind, opposing over-allocation of resources (leading to waste) on one hand and under-allocation on the other (leading to degraded quality).</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*RSoAlrbuYGmumEyWQz7nKg.png" /><figcaption>(over|under)-provisioning</figcaption></figure><p>All of this highlights the inability to align at any given time the allocation (or provisioning if you prefer) of resources with their actual consumption. In short, the opposition between container and content needs to be actively managed and supported in the enterprise.</p><blockquote>The opposition between container and content needs to be actively managed and supported in the enterprise.</blockquote><p>The impact of the Public Cloud, especially the generalization of pay-per-use billing, leads to a new momentum to the handling of this content/container opposition for two main reasons:</p><ul><li><strong>It modifies the granularity of the content</strong>.<br>IT resources deployed are more abundant and diverse, easily manageable and can be assembled together.<br>In short, they can be assigned the 4 properties of the <a href="https://en.wikipedia.org/wiki/Create,_read,_update_and_delete">CRUD</a> acronym (Create Resources, Update and Delete Resources, as well as the constant ability at any time to know their properties and status), previously reserved for data management.</li><li><strong>It improves its optimization</strong>.<br>The savings can be viewed in near real time. The inertia resulting from large CAPEX investments and their amortization over several years disappears.</li></ul><p>For instance, let’s compare the distribution of costs between a legacy on-premise infrastructure (including Private Cloud) and an IaaS Public Cloud with pay-per-use:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*n21pR5i33nUEupwOqmER_w.png" /><figcaption>FinOps today</figcaption></figure><p>Previously, the very high cost of equipment (facilities, energy, consumables, maintenance, etc.) committed the company over several years and had to be renewed at a regular basis by mobilizing significant human resources (managing RFP/Q, technical and legal expertise, purchasing process, etc.). The levers, although existing, were difficult to operate: limited gains whose impact was slow to show up.</p><p>Today, these investments are moving away. The big FinOps challenge is about IaaS resources and their use.</p><p>Currently, the FinOps tools on the market mostly collect usage metrics from the deployed IT resources and give recommendations about the changes to be made. <strong>Right Sizing is king</strong>: it’s all about getting the allocated as close as possible to the consumed.</p><p>If we make the analogy with AWS services, here is the evolution of a typical infrastructure:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Not21eRQNqQjrGvex3YtTA.png" /><figcaption>FinOps today in AWS</figcaption></figure><p>Adjusting EC2 instances, EBS volumes, and RDS databases as needed does not in itself change the container/content dichotomy. We are not far from the waste chase that can also be carried out in an on-prem infrastructure.</p><p>Game changer is that the savings are easier to achieve, quicker to obtain and far from negligible.</p><h4>Evolution to the Serverless</h4><p>To put it bluntly, it is highly likely that Serverless will end this container/content opposition completely. In this new paradigm, resource allocation will either increase or decrease based on usage. In other words, in the analogy of the glass being half empty or twice too full, the glass will grow as the water level rises.</p><blockquote>In the analogy of the glass being half empty or twice too full, the glass will grow as the water level rises.</blockquote><p>FinOps generally speaking (practices, skills, culture, tools) will have to adapt and migrate gradually from an Ops/Infra orientation to a Dev orientation. It will be mostly a question of application efficiency and code optimization.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*hEYAdIyhUp68_r-xd2yYJA.png" /><figcaption>Serverless FinOps in AWS</figcaption></figure><p>Extending the analogy with AWS, the Serverless service pricing model already leaves little room for hunting down wasted resources.</p><p>There are still some signs of this in AWS Lambda (resource allocation, especially the memory allocated to the function — see <a href="https://medium.com/foxintelligence-inside/how-we-reduced-lambda-functions-costs-by-thousands-of-dollars-8279b0a69931">here</a> for FinOps in Lambda) and AWS Aurora Serverless (allocation per ACU slice, i.e. a Memory/Compute combo — see also <a href="https://www.jeremydaly.com/aurora-serverless-the-good-the-bad-and-the-scalable/">this excellent article</a>): it is therefore still possible to fine tune cost drivers related to resource allocation.</p><p>But there is a clear trend towards generalizing consumption to actual usage and abandoning resource allocation/provisioning : we will pay in GB, in IO, in runtime, in number of requests… what our applications will actually consume.</p><blockquote>We will pay in GB, in IO, in runtime, in number of requests… what our applications will actually consume.</blockquote><h3><strong>Welcome to FinDev!</strong></h3><blockquote><strong>FinOps will therefore become a code-centric optimization task devolved to the development teams.</strong></blockquote><p>In a hypothetical « 100% Serverless and On-demand » world, <strong>FinOps will therefore become a code-centric optimization task devolved to the development teams</strong>. Probably with lower gains, as savings are more limited and harder to obtain.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*83_9vy3zViseqIBSwCnLsg.png" /><figcaption>FinOps evolution</figcaption></figure><p>These new <a href="http://finops.world/">FinOps practices</a>, let’s say FinDev to differentiate them, will then have to evolve as well as the tools. New topics will become more important, especially at scale, for instances :</p><ul><li>Which libraries for optimal CPU and memory consumption?</li><li>Shall we consume more CPU and less storage or vice versa for a given Use Case?</li><li>Which data formats to adopt (interesting example <a href="https://cloudforecast.io/blog/using-parquet-on-athena-to-save-money-on-aws/">here</a>)?</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*FV3IpppwmQ6mse4J5yE-5A.png" /><figcaption>DevOpsFin triangle</figcaption></figure><h3><strong>Conclusion</strong></h3><p>But don’t worry, this hypothetical world is far from existing and you won’t have time to get bored! The IT landscape will always remain very heterogeneous: mix of IaaS and Serverless, mix of clouds (multicloud), PaaS and especially SaaS offers, which also differ in terms of financial management.</p><p>Other criteria can — and must — also be taken into account when optimizing expenditure:</p><ul><li>The profile of the applications and their predictability in terms of load (24×7 vs. Business hours, constant demand vs. high variability).</li><li>Purchase options with Reserved Instances, AWS Saving Plans and other Committed Use Discounts over 1 or 3 years</li><li>Service-specific options or resilience commitments</li><li>Lock-in</li></ul><p>Financial studies, Business Cases and other Cost Evaluation still have a bright future ahead of them.</p><p><em>Note: Paper initially published on </em><a href="https://timspirit.fr/avenir-finops/"><em>timspirit.fr</em></a><em>.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=2800a2286c9d" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Muscle up your cybersecurity using Public Cloud]]></title>
            <link>https://medium.com/@a.lagier/muscle-up-your-cybersecurity-using-public-cloud-60263be68870?source=rss-65b5dacc42a6------2</link>
            <guid isPermaLink="false">https://medium.com/p/60263be68870</guid>
            <category><![CDATA[security]]></category>
            <category><![CDATA[cloud-computing]]></category>
            <category><![CDATA[public-cloud]]></category>
            <category><![CDATA[cybersecurity]]></category>
            <dc:creator><![CDATA[Antoine Lagier]]></dc:creator>
            <pubDate>Wed, 10 Apr 2019 11:48:25 GMT</pubDate>
            <atom:updated>2019-04-11T08:56:37.931Z</atom:updated>
            <content:encoded><![CDATA[<p>In early 2019, if you searched for the keyword « Clusif CSV » in bing.com, you gained access to a list of the names of some 2200 IT Security experts (and not the lesser of them).</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/600/1*E2RlxWOBrg-984mmuJHPFg.jpeg" /></figure><p><a href="https://www.01net.com/actualites/les-donnees-personnelles-des-cadors-de-la-cyberdefense-francaise-ont-fuite-sur-la-toile-1631635.html">An ordinary data leak ?</a> Well, yes and no : the funny thing about this leak is that the targeted entity was the CLUSIF, the french Information Security Professionals Club, whose goal is to unite professionals (both vendors and client-side) in the cybersecurity space to exchange ideas, best practices and experiences. I have no doubt what the topic of the next gathering may have been…</p><p>Nevertheless, let’s not poke too much fun at this mishap. Cybersecurity in the digital, ultra-connected, open &amp; mobile world is not an easy task.</p><p><strong>In this context, is the public cloud today an aggravating factor, or, quite the opposite, an ally to better secure your Information Systems and your Data ?</strong></p><p>The reflex of CIOs is still, most of the time, to think that security is stronger “on premises” — when their data &amp; infrastructure is in their own, well known Datacenters. They tend to favor public clouds only for non-sensitive data &amp; applications.</p><p>A page is nevertheless being turned before our eyes : a<a href="https://www.skyhighnetworks.com/cloud-security-blog/microsoft-azure-closes-iaas-adoption-gap-with-amazon-aws/"> recent McAfee study</a> shows that 63% of the IT professionals now consider the Public Cloud as secure as as their own infrastructures.</p><p>At Timspirit, we think that the 37% who argue the opposite are wrong ! Here’s why.</p><h3>Because Cybersecurity requires huge investments</h3><p>Cybersecurity and Data compliance have become so important and complex that it is difficult, even for large organizations, to stay up to date and invest in the massive know-how, systems, tools, knowledge required.</p><p>To highlight this point, look at Google investment in this area :</p><ul><li>After <a href="https://en.wikipedia.org/wiki/Heartbleed">Heartbleed</a> (2014), Google set up a ZeroDay vulnerability discovery team. <a href="https://googleprojectzero.blogspot.com/2014/07/announcing-project-zero.html">Google Project Zero</a>, goes beyond covering Google’s own services &amp; products. It contributes to identify major security vulnerabilities and to correct them before they become <a href="https://blog.lastpass.com/2017/03/security-update-for-the-lastpass-extension.html/">public knowledge</a></li><li>Google has given away no less than $3 million in 2016 in its <a href="https://www.google.com/about/appsecurity/programs-home/">Bug Bounty</a> program, that aims at securing its <a href="https://medium.freecodecamp.org/messing-with-the-google-buganizer-system-for-15-600-in-bounties-58f86cc9f9a5">services</a></li><li>All vendors are investing heavily in research, academia, open source project contributions and to promote the adoption of security standards (such as the generalization of the https protocol)</li></ul><p>Within these large Cloud Services Providers (Microsoft, Google, AWS), <strong>between 600 and 800 experts &amp; professionals</strong> are working daily to guarantee Datacenter &amp; Infrastructure security. Who, even within the largest Fortune 500 companies, can claim to invest as much?</p><p>Think Compliance. Again, it is difficult to achieve certifications for the many, country-dependent certifications that may be required by regulatory authorities. Here’s a quick list of AWS certifications.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/800/1*I3q7wlOdlcclvb1gHUtIvw.png" /></figure><p>Again, how much would it cost to achieve and maintain certification of your own infrastructures up to the same standards?</p><p>Theses means are not enough, of course, to actually guarantee that no security breach will occur in the public clouds. But they offer a strong base layer that is hard to replicate in house — and it would be foolish not to leverage it.</p><h3>Because the public cloud helps clarify your security model</h3><p>The standard security model that has been proposed by a large majority of cloud Providers is a shared responsibility model : the lower levels are taken care of by the provider, the higher levels being the client’s responsibility. Depending on the service model, the threshold gets allocated differently. The higher the outsourced level of the service, the higher the supplier responsibility. Nice, isn’t it ?</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/435/1*xpS50X00z7PtEpVT6AMGtg.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/320/1*6aJNK7M9jxZCwwLE4AUEMA.png" /><figcaption>Shared responsibility model from AWS and Azure</figcaption></figure><p>This model is now evolving into a more robust pattern: the Cloud Service Provider (CSP) are becoming advisors &amp; partners, including for higher levels that remain under the client’s final responsibility.</p><p>The security-related service portfolio offered by these actors has now incomparably more breadth and depth than what can be built in house: logs &amp; access management, audit, configuration management, encryption, flow management, you name it…</p><p>I have witnessed large accounts Chief Information Security Officers decide to use some of these services as a key decision point to move to public clouds, when facing their boards or executive committees.</p><h4>Because the public Cloud helps you build Security deep into your applications — “security as code” or “security by design”</h4><p>Instead of old-fashion « Security Policies » (essentially, word, excel or PDF documents) that all operational teams were supposed to answer to, and which made compliance checks and control a manual, error-prone task, it is now possible to offer embedded, proactive, continuous security compliance control. A few lines of code in a lambda instance allows you to measure real-time compliance with simple rules that are themselves created as code.</p><p>After DevOps &amp; Infra-as-Code, welcome to the brave new world of <strong>Security as Code</strong>.</p><p>Several clients at Timspirit are embarking on this new paradigm shift. They usually use this first as an auditing tool, sending detected non-compliances as events to their Security Operations Teams. They then implement automated remedial strategies for the most common &amp; most critical security controls.</p><p>Get ready to hire a developer as you next CISO !</p><h3>Because you can build a mature &amp; differenciated information security infrastructure</h3><p>An important part of applications and data are usually transferable to Public Cloud « standard » security. For the minority of critical data, public cloud is not excluded however.</p><p>It is possible to secure sensitive data in the Public Cloud, while keeping internal mastery removing any possibility for the provider to access and use these data, even when they reside in their infrastructures. AWS offers Gemalto (purchased by Thales) SafeNet boxes as part of its secure CloudHSM service, allowing customers to manage encryption keys without AWS ever being able to access them. You can also use “Bring Your Own Key” approaches to encrypt data in the cloud.</p><p>We have seen or engaged with military-related organizations, research organizations and Fortune 500 who cracked the solutions repeatedly.</p><h3>Because you can exchange and build solutions with more mature Partners</h3><p>The « Cloud Act » (which forces the American cloud providers, AWS Azure &amp; Google Cloud Platform to supply data to US authorities upon request, within the US criminal law enforcement framework) is still seen as a major risk by European companies.</p><p>Cloud Suppliers are maturing on the topic and become more transparent &amp; helpful to their clients. Microsoft, for example, is engaged in a <a href="https://www.theregister.co.uk/2017/01/24/us_government_microsoft_email_seizure_appeal/">US Lawsuit</a> to counter the FBI’s demand to obtain the emails of non-US citizen, hosted on Irish soil. The same Microsoft has chosen Deutsche Telekom as a trusted third party, covered by German law, to host Azure’s German region.</p><p>If these shows of good will and concern are more and more <a href="https://cloud.google.com/access-transparency/">common</a>, the suppliers will of course still have to abide by the legal framework of their origin countries.</p><p>However, it is now possible to subcontract a large part of operational data/application security to trusted partners, who are better armed and equipped to handle it. And more importantly even, it is now possible to have a transparent, realistic discussion about this with them.</p><h3>Because cyber security complexity is not going to get simpler</h3><p>The frontiers between end-users, networks and servers as we know them is blurring. Think Edge Computing, IoT, or — worse even — the arrival of 5G. These will force CIOs &amp; CISOs to a long overdue aggiornamento of the way they think, design, implement and manage Security.</p><p>Facing this never-ending complexity, it is becoming impossible to cover security on its own.</p><p>Most organizations will be better off partnering with industrial-grade suppliers, whose livelihood clearly depends on their security investments, for base security. They should concentrate their internal resources on higher value tasks — identifying major risk scenario, how to cope with major attacks, survive those risks, and how to spread the security culture. Helping your teams adopt the “move that saves the day” : no external actor can do this for you !</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=60263be68870" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>