Efficient Query Processing in Geographic Web Search Engines

Abstract

Microblogging services became among the foremost fashionable services on the net within the previous
couple of years. This LED to important increase in information size, speed, and applications. This paper presents Venus;
a system that supports period abstraction queries on microblogs. Venus supports its queries on a abstraction boundary R
and a temporal boundary T, from that solely the top-k microblogs area unit came back within the question answer
supported a spatio-temporal ranking operate. Supporting such queries needs Venus to digest many various period
microblogs in main-memory with high rates, yet, it provides low question responses and economical memory utilization.
to the present finish, Venus employs: (1) AN economical in-memory spatio-temporal index that digests high rates of
incoming microblogs in real time, (2) a ascendible question processor that prune the search area, R and T, effectively to
produce low question latency on various things in real time, and (3) a gaggle of memory optimisation techniques that
give system directors with completely different choices to save lots of important memory resources whereas keeping the
question accuracy nearly excellent. Venus memory optimisation techniques build use of the native arrival rates of
microblogs to neatly shed microblogs that area unit sufficiently old to not contribute to any question answer.
additionally, Venus will adaptively, in real time, alter its load shedding supported each the abstraction distribution and
therefore the parameters of incoming question masses. All Venus elements will accommodate completely different
abstraction and temporal ranking functions that area unit ready to capture the importance of every dimension otherwise
counting on the applications needs. in depth experimental results supported real Twitter information and actual locations
of Bing search queries show that Venus supports high arrival rates of up to sixty four K microblogs/second and average
question latency of four unit of time.