|
演讲简介
Big data analytics is often prohibitively costly, and is considered a privilege of big companies that can afford a large cluster of machines. This talk argues that big data analytics is within reach of small companies with limited resources. We present BEAS, a new query evaluation paradigm with constrained resources, based on a theory of bounded evaluation and a data-driven approximation scheme. Better still, BEAS is built on top of existing commercial DBMS and provides small companies with an immediate capacity of querying big data. As a proof of concept, it has been verified that BEAS improves query evaluation of our industry collaborators by orders of magnitude.
大数据分析通常代价高昂,而且被认为是拥有大型集群计算机的大公司的特权。报告认为,大数据分析是资源有限的小公司也可以做到的。我们介绍的 BEAS 系统,就是一种新的有限资源式查询处理范式,它是基于有界估值理论和一个数据驱动近似方案。更好的是,BEAS 是建立在现有的商业数据库管理系统(DBMS)之上,为小公司提供了即时查询大数据的能力。作为概念证明,已经证实 BEAS 可以按数量级提高行业合作伙伴的查询处理。