1. 图规模与CPU平均利用率
真实图
dataset Nodes Edges Hadoop Spark
wiki-Vote 7115 103689 26.81 27.01
soc-Slashdot0902 82168 948464 30.55 31.93
web-Google 875713 5105039 30.82 29.05
cit-Patents 3774768 16518948 29.01 28.37
twitter-Small 11316811 85331845 22.53 30.77
dataset Nodes Edges Hadoop Spark
kronecker19 416962 3206497 31.09 30.67
kronecker20 833566 7054294 30.42 28.03
kronecker21 1665554 15519448 29.62 26.35
kronecker22 3330326 34142787 26.86 26.94
kronecker23 6654956 75114133 23.55 27.37
kronecker24 13305449 165251092 21.98 28.43
预测
dataset Hadoop-Real Hadoop-Predicted Spark-Real Spark-Predicted
soc-Pokec 28.03 29.23 23.49 28.59
soc-LiveJournal 24.4 27.08 27.08 28.65
v23 23.58 25.12 27.58 28.7
v24 22.49 19.9 29.29 28.83
2. 图规模与最大内存占用量
真实图
dataset Nodes Edges Hadoop-System Spark-System Hadoop-App Spark-App
wiki-Vote 7.12 0.1 8835 9980 8112 8835
soc-Slashdot0902 82.17 0.95 9936 16119 8521 14705
web-Google 875.71 5.11 9539 23565 8977 22200
cit-Patents 3774.77 16.52 14679 34689 9537 31901
twitter-Small 11316.81 85.33 39689 71419 10977 65361
模拟图
dataset Nodes Edges Hadoop-System Spark-System Hadoop-App Spark-App
kronecker19 416.96 3.21 10025 22620 8992 20995
kronecker20 833.57 7.05 9779 24392 9362 22793
kronecker21 1665.55 15.52 10778 29207 9475 26537
kronecker22 3330.33 34.14 17568 48849 9744 45647
kronecker23 6654.96 75.11 31789 70917 10492 64606
kronecker24 13305.45 165.25 57540 74061 11655 66702
预测
dataset Hadoop-Real Hadoop-Predicted Spark-Real Spark-Predicted
soc-Pokec 9361 9144.56 31253 26075.74
soc-LiveJournal 10275 9839.54 65295 39760.01
v23 10211 10474.5 65064 52262.6
v24 11845 12164.22 68251 85533.64
3. 图规模与平均磁盘写带宽
真实图
dataset Nodes Edges Hadoop Spark
wiki-Vote 7115 103689 0.43 0.06
soc-Slashdot0902 82168 948464 0.31 0.07
web-Google 875713 5105039 1 0.1
cit-Patents 3774768 16518948 3.41 1.34
twitter-Small 11316811 85331845 6.05 1.56
模拟图
dataset Nodes Edges Hadoop Spark
kronecker19 416962 3206497 0.57 0.08
kronecker20 833566 7054294 1.16 0.29
kronecker21 1665554 15519448 2.77 1.16
kronecker22 3330326 34142787 3.6 1.55
kronecker23 6654956 75114133 5.65 1.69
kronecker24 13305449 165251092 7.55 1.64
预测
dataset Hadoop-Real Hadoop-Predicted Spark-Real Spark-Predicted
soc-Pokec 3.71 1.82 1.44 1.38
soc-LiveJournal 5.27 3.49 1.67 1.61
v23 5.57 5.03 1.69 1.82
v24 7.68 9.1 1.83 2.39
图规模与磁盘读写总量
真实图
dataset Nodes Edges Hadoop Total Read Spark Total Read Hadoop Total Write Spark Total Write
wiki-Vote 7115 103689 959.77 246.46 843.04 10.8
soc-Slashdot0902 82168 948464 473.13 288.53 648.14 14.96
web-Google 875713 5105039 237.95 363.57 2470.1 42.56
cit-Patents 3774768 16518948 702.35 553.43 13187.67 1649.25
twitter-Small 11316811 85331845 1155.43 1432.82 75356.03 5372.9
模拟图
dataset Nodes Edges Hadoop Total Read Spark Total Read Hadoop Total Write Spark Total Write
kronecker19 416962 3206497 221.1 317.43 1261.29 25.51
kronecker20 833566 7054294 247.85 368.12 2981.6 145.34
kronecker21 1665554 15519448 336.35 496.95 9059.13 1046.19
kronecker22 3330326 34142787 533.31 777.34 18707.87 2852.98
kronecker23 6654956 75114133 970.59 1327.62 57725.38 6538.12
kronecker24 13305449 165251092 2012.88 4020.28 163385.83 14736.95
4. 图规模与平均网络I/O带宽
真实图
dataset Nodes Edges Hadoop Spark
wiki-Vote 7115 103689 0.46 0.16
soc-Slashdot0902 82168 948464 0.62 0.49
web-Google 875713 5105039 1.32 1.57
cit-Patents 3774768 16518948 2.62 2.26
twitter-Small 11316811 85331845 3.3 2.07
模拟图
dataset Nodes Edges Hadoop Spark
kronecker19 416962 3206497 0.92 1.3
kronecker20 833566 7054294 1.4 1.77
kronecker21 1665554 15519448 2.13 2.17
kronecker22 3330326 34142787 2.84 2.37
kronecker23 6654956 75114133 3.39 2.43
kronecker24 13305449 165251092 3.62 2.26
预测
dataset Hadoop-Real Hadoop-Predicted Spark-Real Spark-Predicted
soc-Pokec 2.51 3.19 2.34 2.9
soc-LiveJournal 3.17 4.66 2.46 3.39
v23 3.27 5.99 2.5 3.83
v24 3.71 9.55 2.29 5.02
图规模与网络读写总量
真实图
dataset Nodes Edges Hadoop Total Read Spark Total Read Hadoop Total Write Spark Total Write
wiki-Vote 7115 103689 1001.27 29 907.5 28.18
soc-Slashdot0902 82168 948464 1371.84 104.17 1296.8 102.41
web-Google 875713 5105039 3425.08 701.98 3312.13 689.31
cit-Patents 3774768 16518948 10598.61 2917.7 10317.65 2853.13
twitter-Small 11316811 85331845 42737.42 7486.52 41845.98 7296.34
模拟图
dataset Nodes Edges Hadoop Total Read Spark Total Read Hadoop Total Write Spark Total Write
kronecker19 416962 3206497 2158.04 427.08 2082.52 416.83
kronecker20 833566 7054294 3790.94 924.67 3670.59 903.46
kronecker21 1665554 15519448 7325.22 2036.18 7102.39 1997.57
kronecker22 3330326 34142787 15395.38 4565.45 15033.17 4459.93
kronecker23 6654956 75114133 36033.69 9765 35322.16 9572.2
kronecker24 13305449 165251092 81504.24 21223.02 79591.84 20726.19