Search, Power and Energy Saving

posted Jan 31, 2012, 6:40 AM by Jean-Michel Davault

Introduction

Whenever you search across a web application, either a Google/Bing or an internal company one, you expect from a user perspective to get :

-  at least one response,

-  a quick response, subjectively below one second,

-  a powerfull response, in other words « relevant ».

You might have forgotten an energy cost associated to that search. To simplify, computers need electric power to deliver the requested query. The bigger volume needs to be analyzed, the better relevancy is expected, the higher the costs. Here are two indicators :

- in average, energy required to provide a query response is equivalent to one hours from an energy saving bulb,

- just considering a company like Google, their annual energy consumption is estimated between 1900 MW (accounting) to 2500 MW (including their own production), which is more than a nuclear plant.

 

Important changes in progress

Why are you interested?

We have just spoken about « search» which is just considering sending a query and getting results. More and more query suggestions are provided « on the fly » in web enabled applications to “act on data » (discovery). Keep you regularly informed.

Our objective is to deliver a relevant response at a competitive cost within a second. Our value proposition includes either consulting or delivery of secure innovative and robust solutions.

We will present you here on this web site methods and technologies illustrating our expertise and business. Among other, we will publish some posts on :

- green software development,

- Big Data indexation savings,

           - an «associative memory system with enhanced context» …

 
Comments