ENERGY-EFFICIENT QUERY PROCESSINGIN WEBSE ARCHENGINES

Abstract

Hundreds of question handling nodes, i.e. web servers dedicated to fine-tuning consumer
enquiries, make up web search engines. In order to keep response times as low as possible for clients, it
is necessary to have a large number of web servers, which use a large amount of electricity (e.g., 500
ms). In spite of this, clients are unable to uncover response times that are far quicker than those they
had anticipated. Predictive Power Saving Online Organizing Formula (PESOS) is our top recommendation
for selecting the best CPU frequency for fine-tuning a question per core. Prioritize questions by their due
dates and use top-level information to reduce the CPU power consumption of a question processing
node in PESOS. PESOS uses question effectiveness forecasts, which estimate the number of inquiries and
the time it will take to process each one. TREC ClueWeb09B and MSN2006 question logs are used to test
PESOS. With PESOS, an inquiry handling node may reduce CPU power consumption as much as fifty
percent compared with systems operating at maximum CPU core regularity, according to results.
Furthermore, a PESO surpasses even the most sophisticated competitor with a twenty percent power
save, while the rival requires substantial standard adjustment and also may endure in unmanageable
latency violations.