Due to the unstructuredness and the lack of schemas of graphs, such as knowledge graphs, social networks, and RDF graphs, keyword search for querying such graphs has been proposed. As graphs have become voluminous, large-scale distributed processing has attracted much interest from the database research community. While there have been several distributed systems, distributed querying techniques for keyword search are still limited. This paper proposes a novel distributed keyword search system called <inline-formula><tex-math notation="LaTeX">$\mathsf {DKWS}$</tex-math><alternatives><mml:math><mml:mi mathvariant="sans-serif">DKWS</mml:mi></mml:math><inline-graphic xlink:href="jiang-ieq1-3313726.gif"/></alternatives></inline-formula>. First, we present a <italic>monotonic</italic> property with keyword search algorithms that guarantees correct parallelization. Second, we present a keyword search algorithm as monotonic backward and forward search phases. Moreover, we propose new tight bounds for pruning nodes being searched. Third, we propose a <italic>notify-push</italic> paradigm and <inline-formula><tex-math notation="LaTeX">$\mathsf {PINE}$</tex-math><alternatives><mml:math><mml:mi mathvariant="sans-serif">PINE</mml:mi></mml:math><inline-graphic xlink:href="jiang-ieq2-3313726.gif"/></alternatives></inline-formula> <italic>programming model</italic> of <inline-formula><tex-math notation="LaTeX">$\mathsf {DKWS}$</tex-math><alternatives><mml:math><mml:mi mathvariant="sans-serif">DKWS</mml:mi></mml:math><inline-graphic xlink:href="jiang-ieq3-3313726.gif"/></alternatives></inline-formula>. The notify-push paradigm allows <italic>asynchronously</italic> exchanging the upper bounds of matches across the workers and the coordinator in <inline-formula><tex-math notation="LaTeX">$\mathsf {DKWS}$</tex-math><alternatives><mml:math><mml:mi mathvariant="sans-serif">DKWS</mml:mi></mml:math><inline-graphic xlink:href="jiang-ieq4-3313726.gif"/></alternatives></inline-formula>. The <inline-formula><tex-math notation="LaTeX">$\mathsf {PINE}$</tex-math><alternatives><mml:math><mml:mi mathvariant="sans-serif">PINE</mml:mi></mml:math><inline-graphic xlink:href="jiang-ieq5-3313726.gif"/></alternatives></inline-formula> programming model naturally fits keyword search algorithms, as they have distinguished phases, to allow <italic>preemptive</italic> searches to mitigate staleness in a distributed system. Finally, we investigate the performance and effectiveness of <inline-formula><tex-math notation="LaTeX">$\mathsf {DKWS}$</tex-math><alternatives><mml:math><mml:mi mathvariant="sans-serif">DKWS</mml:mi></mml:math><inline-graphic xlink:href="jiang-ieq6-3313726.gif"/></alternatives></inline-formula> through experiments using real-world datasets. We find that <inline-formula><tex-math notation="LaTeX">$\mathsf {DKWS}$</tex-math><alternatives><mml:math><mml:mi mathvariant="sans-serif">DKWS</mml:mi></mml:math><inline-graphic xlink:href="jiang-ieq7-3313726.gif"/></alternatives></inline-formula> is up to two orders of magnitude faster than related techniques, and its communication costs are 7.6 times smaller than those of other techniques.