Sustainability is becoming an increasingly important topic for businesses, with many executives and administrators looking for ways to reduce their carbon footprint and promote an environmentally friendly practice in all business processes, including desktop virtualization.
We also see that it is increasingly common for organizations to include a requirement for providers to demonstrate their sustainability plan when responding to requests for proposal (RFP).
As more and more businesses shift their operations to the cloud, concerns are mounting around the energy consumption associated with public cloud and private data centers. It's important for business leaders to delve into the crucial subject of sustainability concerning VDI and desktop as a service (DaaS). Further, they should examine their delivery models while considering their potential for sustainability from both a hardware and a software perspective.
Understanding the emissions of desktop virtualization
It's worth it to examine some statistics related to carbon footprints to determine key measurables for this discussion.
Although carbon dioxide (CO2) emissions are the primary greenhouse gas emitted from burning fossil fuels, industrial production and land use, it is important to recognize that they are not the only contributors to climate change. Other gases play a significant role in global warming and are combined into a single metric known as CO2e, which stands for carbon dioxide equivalent.
A single laptop alone can result in a carbon footprint of 430 kg CO2e over a period of four years, as reported by Circular Computing. Additionally, according to Microsoft, the carbon footprint of their Studio 2 laptop model is 601 kg CO2e over three years of use, which increases to 749 kg CO2e when adjusted for a five-year period. There is a valuable list that gives an overview of the different vendors and their kilograms of CO2e for different laptops.
Conversely, a thin client might have a carbon footprint as low as 70 kg CO2e over the same time frame and could be even lower depending on the type of thin client and usage. Furthermore, thin clients often have a longer lifespan and some vendors offer options to repurpose old hardware with a thin client OS, extending the useful life of existing equipment even more.
Taking into account the infrastructure, a single physical server in a data center can be responsible for emitting approximately 8,000-9,500 kg CO2e over a four-year period. Calculated 85% of this emission is from usage, while the remaining is associated with manufacturing and transport. This is equivalent to the combined carbon footprint of 15-20 high-end laptops over the same period, although it does not account for the emissions from other components within the data center, such as the network, storage and the facility itself.
Of course, the exact kilograms of CO2e emissions estimate for a physical server depends on many factors such as the energy efficiency of the server, the utilization rate and the energy mix of the data center where it is located.
Regarding public cloud use, various cloud providers have committed to becoming carbon-neutral or even carbon-negative within the next few years. For instance, Google Cloud intends to operate on carbon-free energy by 2030. Similarly, Microsoft Azure has announced its goal of running entirely on renewable energy by 2025 and achieving carbon negativity by 2030. Amazon Web Services has also pledged to run on 100% renewable energy by 2025. Furthermore, research indicates that the compute carbon footprint can be as much as 70% lower when running workloads in the public cloud compared to an enterprise data center, as evidenced in a research paper from Microsoft.
Google has even introduced a new feature to assist IT professionals in making well-informed decisions about selecting environmentally friendly computing resources. Within the Cloud Console location selectors, a green leaf icon is displayed to indicate the regions with the lowest carbon impact (Figure 1).
And all the cloud providers also provide an emissions dashboard so that each organization can view its own emissions.
Test cases for calculating DaaS and VDI emissions
VDI and DaaS offerings can help reduce the amount of electronic waste generated by organizations. Unlike local desktops that need to be frequently replaced, virtual desktops can be easily updated and maintained from a centralized location. This minimizes the need for frequent hardware upgrades and replacements, which contributes to the accumulation of electronic waste. Furthermore, it is crucial to take into account the TCO when assessing the sustainability of VDI, despite the fact that VDI can be more energy-efficient and generate less electronic waste. Nonetheless, VDI might require additional resources for management and maintenance, resulting in higher upfront costs.
Let's examine a hypothetical simplified scenario to explore the contrast more closely. Imagine an organization that runs a VDI environment in a data center, and the end users access it through laptops. Now, imagine the organization contemplating a shift to either a cloud-based VDI or a DaaS service that utilizes thin clients for 5,000 users. How significant would the difference be between these two options across a period of four years?
Private cloud VDI with high-end laptops
- Enterprise data center with 48 physical servers with hyper-converged storage, which would be close to 432,000 kg CO2e, providing VDI with a base of 4 vCPU and 8 GB of memory.
- 5,000 users with laptops would be close to 2,150,000 kg CO2e with an average of 430 kg CO2e per laptop.
- Total of 2,582,000 kg CO2e over a period of four years.
The example provided is overly simplistic and fails to account for other necessary elements in a private data center, but it is the best estimate without getting too deep into highly variable components. These additional elements to consider include built-in support components and functionality such as backup, networking equipment and additional storage requirements.
Public cloud with thin clients
- Microsoft Azure -- 5,000 VDI instances running with associated storage -- where each VM is close to 170 kg CO2e with a base of 4vCPU and 8 GB memory over a period of four years, which would be 850,000 kg CO2e.
- 5,000 users with thin clients would be close to 650,000 kg CO2e, with an average of 130 kg CO2e per client.
- Total of 1,500,000 kg CO2e over a period of four years.
It is of course important to note that depending on which geographical region these workloads are running it will also impact the CO2e output of these workloads. In the example mentioned above, the Azure data center is in Norway.
Another crucial factor to consider when it comes to workloads on the public cloud is the ability to power off resources that are not being utilized. For example, if we assume that the VDI desktop only needs to be available for 10 hours a day instead of running 24/7, this can significantly reduce emissions. In this case, the estimated carbon emissions would be lowered to around 350,000 kg CO2e, which would bring the total down to 1,150,000 kg CO2e.
Although this example might not be applicable to all scenarios because it is a bit narrow, it serves as an illustration of how VDI or DaaS on a public cloud can help lower the overall carbon footprint of an organization.
Key takeaways about VDI and DaaS sustainability
Though both VDI and DaaS offer some sustainability benefits from the hardware and software end, there is a significant difference between the two delivery models. VDI as a service from major vendors such as Citrix and VMware can support various public and private cloud platforms. However, DaaS often only operates on a public cloud platform, such as Microsoft Azure for Windows 365 and Citrix DaaS, and AWS for Amazon WorkSpaces.
DaaS offers some distinct advantages from a sustainability perspective. It is highly scalable and can be rapidly deployed, meaning that businesses can quickly adapt to changing needs and avoid the waste associated with overprovisioning. DaaS also allows for more efficient resource allocation, as users can be provisioned only the resources they need, reducing energy consumption and waste.
However, DaaS does have its own sustainability challenges to consider. Public cloud providers may not be transparent about their energy consumption or sourcing, making it difficult for businesses to fully evaluate their environmental impact. Additionally, because DaaS is delivered over the internet, it is subject to the energy consumption associated with data transmission.
Another aspect to consider is the flexibility of VDI with a cloud-based deployment, which is typically more versatile than most DaaS offerings. With this deployment model, organizations can opt for new infrastructure as it becomes available from cloud providers. Moreover, newer servers are generally more energy-efficient than older ones, leading to lower energy consumption and even lower kilograms of CO2e.
As one might assume, there is no clear-cut answer to which option is more sustainable. Both VDI and DaaS offer unique benefits and challenges when it comes to promoting environmentally friendly practices. However, regardless of whether an organization uses DaaS or VDI, the fact that public cloud providers are committed to running their data centers entirely on renewable energy makes them an even more appealing platform for sustainability-focused organizations.