In the dynamic arena of cloud computing, sustainability and carbon emissions have emerged as focal points. As organizations make strides towards cloud adoption to harness its manifold advantages, the scrutiny on their carbon footprint sharpens. Previously, we examined how a transition to the cloud could serve as a catalyst for your organization's Net Zero ambitions. However, this transition comes with intricacies, chiefly concerning the efficiency of your existing hosting and the carbon footprint of the hardware earmarked for decommissioning.
The quest for optimizing your cloud infrastructure to mitigate emissions commences with a concrete measurement of these emissions. This foundational step is crucial for identifying significant improvement areas and for assessing the effectiveness of your optimization initiatives over time.
A variety of tools are available to measure the carbon emissions associated with your cloud workloads. These tools, furnished by Cloud Service Providers (CSPs) and third-party vendors, encompass distinct features, methodologies, and limitations. In this narrative, we will traverse through these tools, dissecting their offerings to provide a clearer lens through which you can view your cloud's carbon emissions.
Understanding Carbon Measurement
The complexity of Greenhouse Gas (GHG) emissions calculations necessitates a uniform standard to foster meaningful comparisons between enterprises. The GHG protocol emerges as the predominant standard, embraced by a staggering 90% of Fortune 500 companies for emissions measurement and reporting. Although we'll touch on this standard briefly to furnish some context, the Green Software Foundation offers a comprehensive, complimentary training course for a deeper dive.
The GHG Protocol categorizes emissions into three scopes:
Scope 1 encompasses direct emissions from sources under the company's control.
Scope 2 covers electricity indirect emissions from electricity consumption. For instance, GCP's carbon reporting methodology includes scope one emissions from diesel backup generators and scope two emissions from local grid electricity consumption.
Scope 3, encapsulates other indirect GHG emissions not owned or controlled by the company, pivotal in tracking the full hardware lifecycle emissions for cloud providers.
It is commonly noted among Cloud Service Providers that emissions are distributed across the three scopes in a particular fashion: Scope 1 typically encompasses around 10 to 15% of emissions, Scope 2 aligns similarly or slightly higher with about 15 to 20%, while the remainder predominantly falls under Scope 3. The estimation of Scope 3 emissions, particularly those associated with the sourcing and disposal of servers and networking hardware, is crucial. Additionally, Scope 3 emissions pertinent to Data Centers, encompassing aspects like building material and employee commute, hold significant relevance. This nuanced understanding aids in fostering a more comprehensive approach towards cloud sustainability.
It merits attention that for your organization, all emissions emanating from your cloud provider pertinent to your activities are categorized under your Scope 3 emissions. Optimally, we would envisage our tooling to furnish reports on all three scopes to render a holistic view of our carbon footprint. This broader perspective is instrumental in crafting informed strategies towards achieving sustainability benchmarks.
Cloud Service Provider Tools
AWS: Customer Carbon Footprint - CCF Tool
AWS unfurls its Customer Carbon Footprint - CCF tool to all accounts, presenting to all authorized users a vista of emissions over time, segmented by geography and service. However, the geographical segmentation is rather broad, displaying groupings like AMER and EMEA instead of specific AWS Regions. The service breakdown is primarily confined to EC2 and Amazon Simple Storage Service (S3), bundling emissions from other services into a singular figure. This lack of granularity may pose challenges in driving meaningful optimizations. Notably, the data is lagged by three months, and scope three emissions are slated for inclusion in early 2024.
Scope 3 emissions are missing. AWS will communicate their availability in the dashboard by early 2024.
A notable limitation arises in the provision of Scope 2 emissions, which are solely available through a market-based calculation. This implies that AWS refrains from displaying emissions when carbon credits have been purchased, or when investments have been made in projects or renewable energy procurement to counterbalance the emissions. This methodology not only presents an incomplete picture but also falls short of compliance with EU regulations governing carbon accounting and reporting. These regulations mandate the disclosure of GHG emissions under both market-based and location-based methods, thus rendering a more comprehensive and compliant account of emissions.
Given the insufficient granularity of the data, the three-month delay in data refresh, the omission of scope 3 emissions, and the inaccurate/incomplete representation of scope 2 emissions, it becomes challenging to envision how the AWS CCF tool could serve as a reliable means to establish a carbon footprint baseline or spearhead optimization initiatives.
Azure: Impact Emissions Dashboard
Azure's Impact Emissions Dashboard, anchored on Microsoft’s Power BI Pro, delineates emissions by Azure Subscription, Region, and Service. Unlike its AWS counterpart, this dashboard incorporates Scope 3 emissions, albeit access is restricted to customers on an Enterprise Agreement, Microsoft Customer Agreement, or Cloud Solution Provider contract. The methodology for Scope 3 emissions calculation is well-documented, providing a semblance of transparency. However, it appears that scope two emissions adopt a Market-Based approach, potentially skewing the emissions data if renewable energy purchases are prevalent.
GCP: Carbon Footprint
GCP’s Carbon Footprint tool is a readily accessible asset for all accounts, delineating emissions by GCP Project, Region, and Product. It meticulously covers all three scopes of emissions, employing a location-based approach for scope two emissions. An intriguing facet of GCP's approach is the hourly emissions calculation, which considers the fluctuating mix of energy sources in the local grid, thereby enhancing the accuracy of emissions data. Although data is matched hourly, the dashboard is refreshed monthly, offering a balanced blend of precision and manageability.
The Open-Source CCF
Beyond the ambit of CSP-provided tools, third-party solutions like the Open-Source Cloud Carbon Footprint (CCF) offer a consistent approach to emissions measurement across multiple cloud providers.
Originating from Thoughtworks, CCF utilizes a method akin to Etsy’s Cloud Jewels approach, leveraging itemized billing information from AWS, GCP, and Azure to estimate emissions. The tool adeptly utilizes data concerning energy consumption across diverse server hardware, the average emissions correlating to a specified quantum of energy on the local grid, and your itemized usage to calculate estimated emissions for your workloads. Additionally, it ventures to estimate the Scope 3 emissions for server hardware by proportionally allocating the envisaged embodied carbon based on your usage. However, the tool currently falls short in estimating embodied carbon for networking hardware and does not encompass other Scope 3 emissions, to which only cloud providers have access, such as employee commutes.
The tool presents a breakdown by region, account, and service, aiding in identifying emission hotspots warranting attention. The data can be refreshed daily, marking a significantly higher frequency compared to the updates on CSP dashboards. This feature proves beneficial for continuous monitoring and optimization of emissions, as it facilitates prompt detection of any spikes resulting from specific changes, enabling timely rectifications or rollbacks. To bolster optimization efforts, the tool integrates with recommendation APIs provided by CSPs. These APIs adeptly pinpoint areas such as overprovisioned hardware and idle machines. Leveraging this information, CCF can discern and prioritize changes that would yield the most substantial emissions savings.
Boavizta Cloud Scanner
The Boavizta Cloud Scanner is an open-source tool designed to assess the environmental impacts of AWS Instances (EC2) usage by conducting an inventory of your cloud resources and collecting data such as instance types, tags, CPU load, or volume size. The data collected is then used to query the Boavizta API, which returns impact data including metrics like Primary Energy Consumption (PE), Abiotic Resource Depletion Potential (ADP), and Global Warming Potential (GWP).
The Cloud Scanner is flexible in terms of usage; it can be operated from a command line for an immediate view of environmental impacts or used as a metric server to monitor and display real-time impacts in a dashboard, aiding in continuous monitoring and optimization efforts.
The methodology used by Boavizta Cloud Scanner involves calculating the environmental impacts at the device or component level based on the usage configuration. The impacts are computed by multiplying the duration, an impact factor, and electrical consumption. However, the documentation does not explicitly detail the granularity or refresh time for the data.
Boavizta Cloud Scanner's ability to filter estimations by tags enhances attribution to a specific server, environment, application, or service, which could be instrumental in identifying and addressing high-emission areas within your cloud infrastructure.
The documentation did not provide specifics on the granularity or refresh time of the data. For a more precise understanding of these aspects, it might be worthwhile to explore the tool further or reach out to the developers through their GitHub repository.
The landscape of tools for measuring carbon emissions in the cloud is as diverse as it is evolving. Each of the eminent Cloud Service Providers—AWS, Azure, and GCP—has sculpted tools with varying degrees of maturity, robustness, and transparency in methodology. Simultaneously, third-party tools like CCF present a great alternative, especially when benchmarking across multiple cloud environments. The journey towards a reduced carbon footprint in the cloud is a continuous endeavour, and the tool you elect to measure emissions is a linchpin in this voyage.
As you traverse this path, a salient consideration is the comparability of data across different tools, especially if your infrastructure spans multiple Cloud providers. A solution like CCF, with its transparent open-source methodology, offers a vantage point for a more coherent comparison, steering you towards more informed decisions in your cloud sustainability journey.
Navigating the realm of cloud sustainability requires a blend of expertise and tailored solutions. At OptimNow, we specialize in guiding organizations on a green operations (GreenOps) pathway, ensuring a balance between technological advancement and environmental responsibility. Reach out to us for comprehensive Cloud Sustainability and GreenOps guidance that's designed to meet your unique needs. Together, let’s build a sustainable cloud infrastructure that not only meets your operational objectives but also contributes positively to the environment.
Reach out to OptimNow – where sustainable cloud solutions thrive.