<?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 Krupa Suchak on Medium]]></title>
        <description><![CDATA[Stories by Krupa Suchak on Medium]]></description>
        <link>https://medium.com/@khsuchak?source=rss-963f9b7f90c1------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*iJsmwOGhDxtHswYlURIoiw.png</url>
            <title>Stories by Krupa Suchak on Medium</title>
            <link>https://medium.com/@khsuchak?source=rss-963f9b7f90c1------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Sun, 22 Nov 2026 03:20:48 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@khsuchak/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[8 technical ways to reduce data centre power consumption]]></title>
            <link>https://medium.com/@khsuchak/8-technical-ways-to-reduce-data-centre-power-consumption-d39fedaa58b5?source=rss-963f9b7f90c1------2</link>
            <guid isPermaLink="false">https://medium.com/p/d39fedaa58b5</guid>
            <category><![CDATA[energy]]></category>
            <category><![CDATA[data-science]]></category>
            <category><![CDATA[data-centre]]></category>
            <category><![CDATA[sustainability]]></category>
            <category><![CDATA[software-development]]></category>
            <dc:creator><![CDATA[Krupa Suchak]]></dc:creator>
            <pubDate>Wed, 17 May 2023 15:54:16 GMT</pubDate>
            <atom:updated>2023-05-18T08:25:47.945Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*qzb_Fnuh8tazEHWggzAb_w.png" /><figcaption>Live Data Centre</figcaption></figure><p><strong>Krupa Suchak, EMEA Director for Data Centre Transactions, explains how implementing data best practice can make computation more sustainable.</strong></p><p>As high metabolic power consumers utilising 1.1% of global energy, data centres remain top of <a href="https://www.jll.co.uk/en/views/mandatory-data-centre-sustainability-reporting-is-here">regulators agenda</a>, facing increasing pressure to become more sustainable, environmentally friendly and energy efficient. Yet the exponential rise in data and the requirements of training complex AI algorithms such as ChatGPT-3 (1.287 gigawatt hours) means more computation. And this in turn requires more power to fuel data centres and more water to cool down the heat transpired.</p><p>Currently our road towards a net-zero, climate neutral data centre is dominated by design interventions such as free air-cooling, electrical and mechanical engineering solutions. To stay on course to achieve net-zero data centres by 2030, we need to take a holistic approach.</p><p>Here are eight best practices for reduced power consumption that individuals and businesses can adopt to reduce their compute resources and grow sustainably.</p><p>1. <strong>Refactoring Code</strong></p><p>Get in the practice of tracking O(N) Order of N — the number of steps your algorithm must take for N number of datapoints to be processed. This number will shed light on the compute resources used and optimising it will provide performance benefits by reducing runtime. It also improves code quality and agility, decreasing maintenance cost and risk of defects.</p><p><strong>2.</strong> <strong>Energy Aware Programming (Green Coding)</strong></p><p>At times tracking O(N) performance doesn’t necessarily lead energy efficiency. We can intentionally opt for specific techniques, tools, and programming languages. Techniques like recycling and reusing green code, tools like Running Average Power Limit (RAPL) and lightweight languages like Elixir.</p><p><strong>3.</strong> <strong>Compute Cost Tools and Autoscaling</strong></p><p>A key indicator to measuring compute is the costs associated. The higher the cost, the more resources are required. There are multiple tools provided by various cloud services, offering levers to control computation and cost. A job can be optimised at low-level before pushing through to production or scaling up. Best practices involve validating hypotheses as proof of concept, iterating on the algorithm to optimise efficiency and runtime and then deploying the model to scale.</p><p>4. <strong>Compute on Edge</strong></p><p>When and where feasible, compute closer to the point of consumption and distribution of data, rather than duplicating and transferring data to a data lake or warehouse. Further to this, run machine learning (ML) models on an edge device like a smartphone to collect, process, and recognize patterns within collections of raw data, in close to real time.</p><p>5. <strong>Virtualisation</strong></p><p>This<strong> </strong>enables users to access hardware services to compute over a distributed network. The hardware is open to multiple users and observes economies of scale for compute purposes. Similar to the concept of saving resources by migrating to the cloud.</p><p>6. <strong>Data Mesh</strong></p><p>Data management is a major, if not the biggest expense in cloud costs. Data mesh is a concept that offers a platform to <em>connect</em> not <em>collect </em>data through a federated decentralised governance. The raw data resides at the place of origin — avoiding duplication — but is integrated for efficiency and simplicity across the architecture.</p><p>7. <strong>Modernisation of Database Architecture</strong></p><p>Instead of lifting and shifting all data when migrating to cloud, consider moving away from relational databases to a more intuitive architecture. Innovations in architectural design are addressing both volume and speed issues of data transfer and computation, swiftly automating job scheduling and enabling parallel processing to reduce computational overheads.</p><p><strong>8.</strong> <strong>Adoption of Latest Technology</strong></p><p>The latest high tech server racks, processors, microchips,<strong> </strong>storage devices and cooling systems are more energy efficient and thus significantly reduce energy consumption in the day-to-day operation of a data centre. As we start delegating more compute tasks to high-tech servers, let’s not forget to move low compute, storage data to redundant data centres with poor compute capacity and retire legacy data centres altogether.</p><p>In the next decade, we foresee data centre consumption values mushrooming, underpinned by the exponential growth of data. However, collectively through adoption on technical approaches we can inch towards our goal of achieving sustainability in data centres by 2030.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=d39fedaa58b5" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>