Telco Cloud

Placing Telecom Workloads to Minimize Energy Consumption and Reduce Carbon Emissions

Telecommunications and IT together, known as ICT, consume 5-9% percent of the world’s electricity supply, and the rapid growth in digitization could increase the information and communications technology industry’s power consumption to 20 percent by 2030. 

This projected increase in energy consumption is driven not by losses in efficiency but by the ever-increasing demand for transmitting and consuming more data at maximum speed. Without a strategy for tackling energy consumption, the projected increase in energy consumption to 20 percent is likely to come true. And then there are carbon emissions. The Intergovernmental Panel on Climate Change (IPCC) recently issued a report based on more than 14,000 scientific research papers that forecasts the dire consequences of failing to control greenhouse gas emissions — without progress toward carbon neutrality, extreme weather, habitat destruction, glacial melting, and rising sea levels will dramatically accelerate. The stakes are high. 

At the same time, mobile networks are undergoing a transformation that increasingly taps complete virtualization and containerization to serve up 5G. What started in the core is rapidly spreading to radio access networks and the edge, with software-defined core, RAN, and edge networks running on systems like VMware Telco Cloud Platform. With more network functionality being moved into data centers or to the edge, including virtualized RAN sites, the share of network energy consumption is projected to grow nearly five-fold. 

Figure 1: Percent of total power consumption that the various parts of a 4G/5G mobile network consume, as presented in “The Path Toward Sustainable 6G

Working smarter to reduce emissions over time 

The telecommunications and IT industries have a responsibility to minimize their emissions and to deliver innovative ways to help our customers minimize their energy consumption and emissions. In addition to virtualization and containerization, smart grids, grid-interactive data centers, and micro-grids–where devices can share and trade low levels of energy to prevent the need for batteries–can all help reduce emissions over time. If it works smarter and manages workloads better, some sources estimate that the ICT industry can prevent emissions at a rate of 10 times its own footprint by 2030

How can we better manage workloads to help minimize carbon emissions? A critical component, of course, is using less carbon-intensive electricity.  

Strategies to minimize energy consumption and reduce emissions 

So, what then are some strategies for minimizing energy consumption and emissions for the telecommunications workloads of the future? Some obvious answers spring to mind: operating energy-efficient hardware, designing compute-efficient applications, and maximizing host utilization, including with virtualization. 

But there are several less obvious strategies to directly help minimize energy consumption and reduce carbon emissions, as our Director of Sustainability Innovation at VMware, makes clear in a recent blog, “Pathways to Sustainable Clouds”:  

  • Making energy consumption visible and making workloads carbon-aware 
  • Using renewable-energy-powered data centers 
  • Placing and scheduling workloads with the explicit, measurable intention of minimizing energy consumption and reducing emissions 

Workload placement and scheduling to control emissions 

The last point is particularly important because if you can manage when and where workloads are running, you can take advantage of renewable energy. A prerequisite to doing so is integrating energy and carbon metrics with your workload management system. The development of carbon-aware software can integrate electricity carbon intensity as an optimization factor into workload management. Intelligent workload placement and optimization can help you chart a course to sustainable 6G. Carbon emissions can vary from near-zero for wind, solar, hydro, and nuclear power plants to carbon-intensive sources like coal and natural gas power plants. The sources that contribute energy to local energy grids typically vary with time and place. 

The good news, though, is that the share of renewable energy and low-carbon electricity reached almost 55 percent in 2019 for global electricity generation. In aggregate, workload placement and scheduling could help reduce demand for carbon-intensive electricity.  

Intelligent workload placement in the grids of the future 

While low-carbon electricity sources are on the rise, the internet is evolving to pave a path for a grid that virtualizes and stitches together edge resources across multiple clouds and locations. The Open Grid Alliance aims for a globally distributed multi-dimensional fabric of heterogeneous compute, data, and intelligence to power on-demand, context-aware, immersive, and distributed modern applications anywhere.  

VMware and Vapor IO co-founded the Open Grid Alliance to enhance collaboration and open architectures by scaling the edge horizontally and expanding small edge data centers. The Multi-Cloud Service Grid (MCSG) is a multi-cloud framework that lets you run hypercomposed applications — applications that are dynamically composed and optimized with the right type and amount of resources at the right time and place. 

A multi-cloud services exchange can enable you to create and operate hypercomposed applications through the following components: 

    1. Data center without borders – distributed abstraction overlay that combines heterogeneous data center elements into a logical pool of resources.  

    2. Distributed Control Plane – MCSG can distribute the control plane and application wherever it makes the most sense. 

    3. Adaptive policy engine – AI-Ops for resilient multi-site operation of distributed applications.  

Together, these services can enable you to place workloads where they minimize energy consumption and carbon emissions.  

Vapor IO’s Kinetic Grid’s real-time intelligence system, Synse, provides critical telemetry to VMware’s Adaptive Policy engine, which analyzes the data to deliver self-adapting optimizations for workload placement and high availability.  

Optimizations to minimize carbon emissions 

Crucially, this intelligent optimization could support goals like meeting application SLAs while minimizing carbon emissions.  

Here is an example of our vision for the future: suppose the Kinetic Grid platform shows two sites, Dallas and Atlanta, as operational. Now, imagine that the Dallas site encounters a problem. Using telemetry data from the Dallas site, the Kinetic Grid shows that the supply of renewable energy has decreased. The adaptive engine from VMware can detect a rise in carbon emissions by communicating with the Synse API at the Dallas site. Fortunately, the Atlanta site reports that it has an excellent supply of renewable energy.  

When the emissions rise at the Dallas site, VMware Telco Cloud Operations receives an alarm notification. The adaptive engine immediately sends a notification to VMware Telco Cloud Automation to trigger the workload migration, which relocates the Kubernetes pods from Dallas to Atlanta. And the Kinetic Grid visually shows your administrators that the workload was migrated successfully from Dallas to Atlanta. 

The result of tapping this concept of a data center without borders and hypercomposed on-demand grid-services is a reduction in carbon emissions through intelligent, automated workload migration. 

Figure 2: The top plot shows the recent carbon intensity of the electricity supplied to the two regions–the pink line is Atlanta, the green is Dallas. We see a number of times when one region has a higher intensity than the other. The bottom plot shows the benefits of carbon-aware workload placement. Pink is the emissions for ’business as usual,’ and green shows the emissions if we shift workloads between data centers to minimize carbon emissions.

Check out our demo on workload placement for Sustainability at MWC Barcelona 2022   

Attending MWC Barcelona 2022? Interested in finding out more about the benefits of building a virtual data center with the VMware Multi-Cloud Service Grid?  

Head over to the VMware booth and check out our demonstration of how VMware Telco Cloud Platform turns infrastructure telemetry into end-to-end AIOps for the data center to perform context-aware resource allocation. It might just spur you on to do your part to minimize the carbon emissions of your telecommunications workloads. The title of the demo is Accelerating Sustainability with Intelligent Workload Placement and Migration.  

Our exhibition at MWC Barcelona 2022 is located at Hall 3 Stand 3M11, Fira Gran Via; see  

Come meet up with us in Barcelona


Leave a Reply

Your email address will not be published. Required fields are marked *