LOAD DISPERSING METHOD FOR USE IN CLOUD STORAGE PERMITTING RANGE SEARCHES

PROBLEM TO BE SOLVED: To propose a load dispersing method for use of range searches required in search for key value type data in cloud storage.SOLUTION: For the quantity of data accumulated in each virtual node, a lower limit and an upper limit barely surpassing the twofold of the lower limit are p...

Full description

Saved in:
Bibliographic Details
Main Authors JIBIKI MASAHIRO, SHAO XUN, TERANISHI YUICHI
Format Patent
LanguageEnglish
Japanese
Published 06.04.2017
Subjects
Online AccessGet full text

Cover

Loading…
More Information
Summary:PROBLEM TO BE SOLVED: To propose a load dispersing method for use of range searches required in search for key value type data in cloud storage.SOLUTION: For the quantity of data accumulated in each virtual node, a lower limit and an upper limit barely surpassing the twofold of the lower limit are prescribed to disperse the load. (1) If addition of data to a first node would result in surpassing of the upper limit, an empty node is used as the second node, and the quantity of data in excess of the lower limit is shared from the first to the second node to update the array of key value type data. (2) If the sum of the data quantities of the first node and the second node is not greater than the upper limit, data is collected into one of the two nodes to make the other an empty virtual node. (3) If the sum of the data quantities of the adjoining first and second nodes surpasses the upper limit and the data quantity of one node is not greater than the lower limit, data is shifted from the other to the one node for redistribution to cause the data quantity of each of both nodes to surpass the lower limit for updating of the array of key value type data.SELECTED DRAWING: Figure 4 【課題】キーバリュー型データについて、データを検索する際に求められる、範囲検索可能なクラウドストレージにおける負荷分散方法を提案する。【解決手段】各仮想ノードのデータ蓄積量に下限と該下限の2倍を僅かに超える上限とを設け、負荷分散を図る。(1)第1ノードにデータを追加すると上限を超える場合は、空ノードを第2ノードとして、第1から第2ノードに下限を超える量のデータを分け、キーバリュー型データの並びを更新し、(2)第1ノードと第2ノードのデータ量の和が上記上限以下の場合に、その一方にデータを寄せて他方を空仮想ノードとし、キーバリュー型データの並びを更新し、(3)隣接する第1と第2ノードのデータ量の和が上記上限を超え、且つ、一方のノードが上記下限以下の場合に、他方から一方のノードへデータを移動して双方が上記下限を超える様に再配分し、上記キーバリュー型データを更新する。【選択図】図4
Bibliography:Application Number: JP20150194688