Power Consumption Testing

The case for going "green" and reducing power consumption within IT is not simply a means of paying lip service to the environmentally aware. It is an absolute requirement. The reality is that companies in high-concentration areas for IT service delivery such as the Docklands in London and other major European cities are hitting the limits of what the energy provider can offer them.   This is forcing enterprise to maximise the use of their existing power and measure usage.

Vendors own published power consumption figures are meaningless - we have found them to be very inaccurate during internal tests in our labs.  By offering accurate power consumption testing Cavell, in partnership with Broadband Testing Labs, can then produce a report with defined metrics and recommendations to increase power efficiency and then decrease or maintain power usage.

Our power consumption testing service, delivered with our partner BroadBand Testing Labs, is designed to generate real-world usage figures based on a per device basis that, by definition, can be aggregated to show overall consumption. This will deliver an overall network MB/kW, a per device MB/kW as well as recommendations on how to advance.

The three main outcomes will be:

  • How to make the most of existing kit – tweaks and changes that can be done to decrease power usage and increase efficiency
  • Modular upgrade recommendations that can be made to make existing kit more effective
  • Upgrade of existing kit if no other effective method can be found

An important starting point is that we have found vendors' own quoted figures to be wildly inaccurate and with no obvious point of source (so what DID they measure exactly)? Using dedicated network test equipment, we are able to measure power consumption in real-world conditions with variable traffic loads and profiles, from idle to completely saturated. In this way we are able to pinpoint absolute power consumption from a combination of devices on the network at the real traffic levels witnessed every day.

Moreover, the test methodology we are using is being proposed as the industry standard methodology in the networking industry, so are truly meaningful and comparable. Based around the existing industry standard RFC2544 methodology for performance testing, the new standard adds in power consumption measurements into this real-world test scenario.

Optimising for power

What are the simple techniques that you can use to optimise your power consumption?  There are four main elements, which combined or used individually, can save you cost and increase your network capacity.

  • Server and computing
    • Server virtualization, VMware, ESX
    • Virtual LAN, VMDq, VMware NetQueue and V10GB interfaces
  • Storage
  • LAN and building
    • Power and cooling impact of e.g. VOIP deployment
    • Power over Ethernet (POE) - from power to the individual desktop towards centralised power
    • Power optimized equipment, decisions for central POE or distributed POE provisioning
  • WAN
    • Power impact of a single converged services network
    • Availability becomes critical, and more redundancy can mean more power consumption
    • Restructuring, optimizing and removing duplication

Quantifying and measuring power also requires a deep understanding of networks, this combination is what Cavell can bring to your business.

For more information or to get a quote please email sales@cavellgroup.com