Discover how network sustainability analytics from HPE, part of Juniper Data Center Assurance helps data centers track energy consumption, optimize power, and report emissions with AI forecasts.
As organizations expand data center infrastructures to support AI workloads, energy consumption has become a critical operational concern. Rising power demand increases operational expenditures, and regulatory frameworks are tightening requirements around emissions reporting and energy efficiency. Organizations need better ways to monitor, optimize, and report on energy usage.
HPE network sustainability analytics, a new Juniper Data Center Assurance feature, provides granular visibility into energy consumption, greenhouse gas emissions, and power cost at the organization, site, and device levels—paired with prescriptive optimization recommendations and AI-based forecasting for SBTi planning and compliance reporting.
Organization-level dashboard: Consolidated view of energy performance
The organization-level dashboard aggregates energy telemetry across all sites. Metric cards display real-time power usage, energy consumption, GHG emissions, and estimated cost. Historical trend charts track consumption, emissions, and cost over time.
- Top sites ranking: Sites listed by energy consumption with sortable columns for cost and GHG emissions, so teams can quickly prioritize the highest-consumption locations.
- Geographic map view: An interactive map overlays energy metrics by site, with circle sizes indicating relative consumption. Hover over any site to see energy data for the past hour.
- Potential energy savings projections: Estimates energy savings, GHG reduction, and cost savings achievable if all device-level recommendations are applied. Projections available for today, tomorrow, next 7 days, and next 30 days, with breakdowns by recommendation type, device model, or Junos OS version.
- Configurable carbon intensity and power cost: Administrators set carbon intensity (gCO₂e/kWh) and average power cost per kWh, feeding all downstream GHG and cost calculations. Site-level overrides are supported.
Figure 1. Organization-level dashboard
Site-level insights: Efficiency analysis across the fabric
The site-level dashboard aggregates device telemetry into three tabs: Overview, Topology, and Potential Energy Savings.
The Overview tab includes core metric cards with trend graphs, plus site-specific analytical tools:
- Power vs. throughput scatter graph: Plots each device by throughput versus power consumption. Devices in the top-left quadrant (low throughput, high power) are underutilized candidates for investigation.
- Power vs. throughput line graph: Shows aggregated site-level power and throughput over time, revealing idle power draw during off-peak hours when throughput drops but power stays flat due to always-on components.
- Top devices table: Ranks devices by energy consumption, with sortable columns for max/min traffic, average fan speed, temperature deviation, and potential energy savings.
The Topology tab visualizes the network fabric (super spine, spine, leaf) with a metric selector overlaying power, emissions, or temperature using a color gradient—helping correlate energy behavior with device roles and identify hotspots.
The Potential Energy Savings tab breaks down estimated savings by recommendation type, device model, and Junos OS version.
Figure 2. Site-level view
Device-level analytics: Granular energy management
The device-level dashboard is organized into four tabs: Overview, Power Supply, Thermal Profile, and Recommendation.
Figure 3. Device-level view Thermal Profile
Overview tab—Metric cards show real-time power draw, energy consumption, GHG emissions, and cost with trend graphs. Additional tools include:
- Power vs. throughput line graph: Shows the historical correlation between power draw and traffic load, exposing the gap between actual demand and constant power draw.
- Per port peak/off-peak traffic heat map: A 24-hour rolling heatmap showing hourly traffic intensity for every port, color-coded by configurable utilization thresholds. Enable the Forecast checkbox to display predicted traffic patterns for the next 24 hours. Use this to identify unused ports, consistently low-traffic ports, and off-peak windows where ports can be safely shut down.
Power supply tab—A progress bar shows actual draw versus maximum capacity. A donut chart summarizes online versus offline power modules, and a table lists each module’s load, status, and conversion efficiency.
Thermal profile tab—Temperature and fan speed time-series graphs detect mismatches, such as fan speed staying high while temperature is low. A power trend chart correlates power draw with thermal load. A top sensors table lists the five sensors with the highest temperature deviation from baseline, and a fan status view shows individual fans within each tray with RPM and health indicators.
Energy optimization recommendations
The Recommendation tab provides prescriptive, device-level optimization guidance. Each recommendation includes a description, configlet-based actions (applied through Apstra Data Center Director), quantified benefits (power savings, cost reduction, GHG reduction), and documented risks. The service runs daily using historical telemetry and configuration data.
- Unused port shutdown—Identifies ports that are administratively enabled but carry no traffic (link-state down, no Tx/Rx activity, or transceiver inserted with zero throughput) and recommends disabling them. Requires at least 24 hours of data.
- Low traffic port optimization—Analyzes historical and AI-forecasted per-port traffic to identify ports that consistently operate below the configured low-traffic threshold during recurring off-peak windows, and recommends temporary shutdown during those periods. Historical analysis requires at least 3 days of data; forecast-based analysis may require up to 15 days.
- Unused line card and PFE shutdown—For modular platforms (Juniper PTX10016 Router, Juniper PTX10008 Router, Juniper PTX10004 Router, Juniper PTX10001-36MR Router), identifies line cards and packet forwarding engines where all associated ports show no traffic and no active links, and recommends powering them down—among the highest-impact actions for reducing chassis-level power consumption and cooling overhead.
Port utilization thresholds that drive the low-traffic recommendations are configurable at both organization and site levels, with site-level settings overriding defaults.
Figure 4. Device energy optimization recommendation
Sustainability report: actual and forecasted performance for SBTi planning
The sustainability report consolidates actual energy usage with AI-based forecasts into a month-by-month view at both organization and site levels. Metric cards display cumulative values for energy consumption, GHG emissions, and cost. Trend charts show each month as a horizontal bar, with historical data in grey and forecasted data in blue.
The report supports flexible time range selection—this year, last year, next year, or custom range—enabling year-over-year comparison and alignment with internal reporting cycles. Reviewed together, the organization and site reports identify whether corrective actions are needed locally or across the organization, and whether performance is tracking toward SBTi goals.
Figure 5. Sustainability yearly report for GHG Emission
Getting started
To learn how network sustainability analytics fits into your sustainability strategy, explore Juniper Data Center Assurance. For broader HPE sustainability analytics across compute and storage, see HPE Sustainability Insight Center and HPE Complete Care Service—Energy and Emissions Reporting.
Start turning your network's idle watts into real savings—explore Network Juniper Data Center Assurance
Meet the author:
Marc Rapoport
Director Product Management
Network Products & Advanced Technology (NPAT)
HPE Networking
