### Table 1: Comparison of shuffle arguments

2003

"... In PAGE 23: ...Table 1: Comparison of shuffle arguments Table1 should of course be read with care. More important than the number of exponentiations is what happens when we throw in randomization, batching and multi-exponentiation techniques.... ..."

Cited by 31

### Table 1: Comparison of shuffle arguments

2003

"... In PAGE 23: ...Table 1: Comparison of shuffle arguments Table1 should of course be read with care. More important than the number of exponentiations is what happens when we throw in randomization, batching and multi-exponentiation techniques.... ..."

Cited by 31

### Table 8. Asymptotic upper bounds for segmentintersection searching, with polylogarithmic factors omitted.

"... In PAGE 42: ... Agarwal and Je Erickson example of segmentintersection-searching in the beginning of this section. See Table8 for some of the known results on segment intersection searching. For the sake of clarity, we have omitted polylogarithmic factors from the query-search time whenever it is of the form n=m .... ..."

### Table 2. Comparison of shuffle arguments for Paillier encryption.

2003

"... In PAGE 30: ... Recently Groth and Lu [27] have suggested a shuffle argument based on homomorphic integer commitments as well as one that uses ideas from Furukawa [18]; we include the latter scheme in the table. In Table2 we compare the arguments for correctness of a shuffle of Paillier ci- phertexts. The parameters we have chosen are a 1024-bit Paillier modulus, which gives 2048-bit ciphertexts, 160-bit challenges and for statistical hiding we use lscripts = 80.... ..."

Cited by 31

### Table 3: The cost of data shuffling with and without foreground OLTP workloads. The no shuffling column represents the average response time during this period if there is no back- ground shuffling.

2005

"... In PAGE 8: ...Table3 in Section 5 shows that our estimates of average response time and phase length are fairly accurate, compared to the experi- mental measurements. The performance model described in Section 3.... In PAGE 11: ... The random shuffling algorithm was used to ensure that future loads could be evenly distributed across the 23 full-speed disks. Table3 shows that the average response time of foreground re- quests only increases by 2.... ..."

Cited by 21

### Table 3: The cost of data shuffling with and without foreground OLTP workloads. The no shuffling column represents the average response time during this period if there is no back- ground shuffling.

2005

"... In PAGE 8: ...Table3 in Section 5 shows that our estimates of average response time and phase length are fairly accurate, compared to the experi- mental measurements. The performance model described in Section 3.... In PAGE 11: ... The random shuffling algorithm was used to ensure that future loads could be evenly distributed across the 23 full-speed disks. Table3 shows that the average response time of foreground re- quests only increases by 2.... ..."

Cited by 21

### Table 8. Asymptotic upper bounds for segment intersection searching, with polylogarithmic terms omitted.

1999

"... In PAGE 38: ... We have already given an example of segment intersection-searching in the beginning of this section. See Table8 for some of the known results on segment intersection searching. For the sake of clarity, we have omitted polylogarithmic factors from the query-search time whenever it is of the form n=m .... ..."

Cited by 205

### Table 8. Asymptotic upper bounds for segment intersection searching, with polylogarithmic factors omitted.

1999

"... In PAGE 42: ... Agarwal and Je Erickson example of segment intersection-searching in the beginning of this section. See Table8 for some of the known results on segment intersection searching. For the sake of clarity, we have omitted polylogarithmic factors from the query-search time whenever it is of the form n=m .... ..."

Cited by 205

### Table 9. Asymptotic upper bounds for ray shooting queries, with polylogarithmic factors omitted.

1999

"... In PAGE 44: ... Following a similar approach, Pocchiola and Vegter [241] showed that a ray-shooting query amid a set P of s disjoint convex polygons, with a total of n vertices, can be answered in O(log n) time, using O(n + m) space, where m = O(s2) is the size of the visibility graph of P.8 Table9 gives a summary of known ray-shooting results. For the sake of clarity, we have omitted polylogarithmic factors from query times of the form n=m .... In PAGE 44: ... The ray-shooting structures for d-dimensional convex polyhedra by Matou sek and Schwarzkopf [205] assume that the source point of the query ray lies inside the polytope. All the ray-shooting data structures mentioned in Table9 can be dynamized at a cost of polylogarithmic or n quot; factor 8The vertices of the visibility graph are the vertices of the polygons. Besides the polygon edges, there is an edge in the graph between two vertices vi; vj of convex polygons Pi and Pj if the line segment vivj does... ..."

Cited by 205

### Table 3. Ensemble on Real Data Shuffling

"... In PAGE 8: ... Also, as illus- trated in Section 1, both P(x) and P(y|x) undergo contin- uous and significant changes in this stream data. Results of various methods on streams generated by Shuffling and Stratified Sampling are summarized in Table3 where margin tolerance rate is set to be 0.001.... ..."