Energy Monitoring Conclusion

This section will reflect on the state of energy monitoring within the IT sector. Evaluate the energy monitoring tool produced as part of this project and look at areas for future technical and commercial development/enhancements.

Summary of the Industry

The IT sector is growing and its appetite for energy now exceeds most countries in the world. Energy prices have risen and are predicted to continue to rise. The earth is getting hotter and IPCC report a causal link between greenhouse gas emissions and global warming.

Energy monitoring products within the IT industry fall into two broad categories, building/estate management and device monitoring. Building and estate energy reporting systems are used widely within the facilities management sector, these systems tend to focus on HVAC systems. Organisations who only monitor HVAC exclude IT equipment energy usage. When organisations monitor energy use on a device, information collated is often used as a proxy for measuring general system health for example, Firewall power and CPU utilisation is often used as a proxy for unusual network traffic.

Companies which have already developed energy monitoring tools which can monitor devices have either extended the product to include management or have focused the product on specific technology types. This project has identified that there is a gap in the market, there are no software solutions which monitor energy for all IP connected devices within a network. At present the only way to monitor all IP devices within a network and report on their individual power consumption is through the physical installation of individual energy probes between the power source and the IP device.

Product Reflection

This project has developed an energy monitoring solution which could be installed in enterprise networks and monitor energy use for all IP devices. The solution collates individual energy usage information without the need for software installed on the monitored device and without the need for additional physical energy monitors. The solution developed produces energy usage reports for all devices individually and also groups this information based on location and device type. The solution provides organisations with a view of IT assets within their estates and report on energy used.

Whilst the product has been proved to be able to measure energy consumption on IP connected devices within a network, key questions remain:-

  • Is this product useful/desirable within the market place?
    • Based on research conducted 60% of respondents currently monitored energy consumption, with fewer than 2 in 10 metering electricity on a 30 minute basis. These responses could indicate that basic electricity monitoring via bills and meters used. The product developed within this project would enhance existing reporting systems with near real time energy monitoring.
  • How accurate and real time does monitoring need to be?
    • Based on testing within the laboratory it was established that the reporting tool provided a 30 second delay between energy usage and display. Based on responses from the questionnaire conducted respondents wanted grouped energy usage reports more than individual energy use per device. An inference of this could be that reporting does need to be accurate but not real time. A 30 second delayed energy usage monitor would be acceptable.
  • How scalable is the solution currently and how scalable does it need to be?
    • The laboratory environment proved functionality and scaled to 37 devices. Within a live customer environment scalability may be a concern. Based on responses from the questionnaire conducted approximately half of respondents represented organisations with less than 250 employees. The product was able to comfortably accommodate polling 37 devices. Based on testing so far, with the current deployment model the platform would be able to poll up to 1024 devices so would be able to support SME organisations.

Product Development

The product could be developed further, areas for additional functionality and improvements are: –

  • Support application programming interface (API) integration – currently the tool is a black box, polling, recording and reporting. Going forward, it would be useful to allow other reporting tools to obtain power usage information and display that information within existing reporting platforms.
  • Scalability – test the platform beyond 1024 active monitored devices. Once the upper limit of the platform is identified, develop the product to monitor additional nodes. This could be achieved by distributing the polling agent to several processes or VMs.
  • Common device settings repository – at present the product relies on manual intervention to support power monitoring of new device types. Going forward it would be useful to use a Git repository to store and distribute common settings files. This would allow for automatic updating of the settings files.
  • Support for IP v61. the product uses IP v4. IP v6 launched in 2012 and is gradually being adopted by organisations.

Future of the Product

Through these posts and we now have an opportunity to trial this solution within a real life customer. The “customer” will allow us to install a stand alone device in their premises and measure their energy use. This free trial will allow us to establish what additional development is needed.

Looking into the distant horizon (next 12 months or so) the intention is to do two things. Launch the product as a beta release and agree on a charging model for the solution. The beta release will be the next phase of software testing in which a live customer trials the product. This release will be installed for a time limited free trial. The aim of this release is to support further testing and put in place any improvements required.

To quote Milton Friedman, “The most important single central fact about a free market is that no exchange takes place unless both parties benefit”2.

The ambition for this product is, it reaches the market, is desirable and is useful. Product demand from the market (i.e. paying customers) will provide a test of how desirable the product is. Customers will only install this solution if, it is of benefit, i.e. energy savings or useful information obtained.

  1. Determinants of IP Version 6 Adoption, (2006). 2006 International Multi-Conference on Computing in the Global Information Technology – (ICCGI’06) Available at: \href{https://doi-org.ezproxy1.bath.ac.uk/10.1109/ICCGI.2006.25}{https://doi-org.ezproxy1.bath.ac.uk/10.1109/ICCGI.2006.25} [Accessed 10 June 2018] []
  2. Milton Friedman Interview Transcript 10th Jan 2000 [Online] Available at:\ \href{http://www.pbs.org/wgbh/commandingheights/shared/minitext/int_miltonfriedman.html}{http://www.pbs.org/wgbh/commandingheights/shared/minitext/int_miltonfried- man.html} [Accessed 28 August 2018] []