Cloud　Computing　promises　to　deliver　on　an　objective　that　building　on　compute　and　storage　virtualization　technologies,　consumers　are　able　to　rent　infrastructure　＇in　the　Cloud＇　as　needed,　deploy　applications　and　store　data,　and　access　them　via　Web　protocols　on　a　pay-per-use　basis.　For　application　and　data　deployment　in　cloud,　high　availability　and　service　level　agreement　must　be　satisfied.　To　leverage　the　performance　and　the　number　of　concurrent　access　of　application　and　data,　dynamic　application　deployment　and　data　replication　are　usually　employed.　The　consistence　of　multiple　data　replicas　is　discussed　in　this　paper　and　lazy　update　approaches　are　used　to　separate　the　process　of　data　replication　and　data　access　in　cloud　computing,　which　improves　throughput　of　the　data　access　and　reduces　response　time.　Initial　experiments　show　the　methods　are　efficient　and　effective.　©2010　IEEE.