Let Q = {91;:::;44} be a set of queries, A = {A1; : : : ;A4} be a set of attributes, and S = {S1,S2,S3,S4) be a set of sites. The matrix of Figure 3(a) describes the attribute usage values and the matrix of Figure 3(b) gives the application access frequencies. Now, a) Calculated attribute affinity matrix for the above scenario. b) Suppose attribute A1, A2 is already placed in clustered affinity matrix. Now place A3 in the clustered affinity matrix. 51 52 S3 54 , , . , 1 0 1 0 91 41 9 11 0 10 e 72 0 1 1 0 8 12 19 0 0 1 0 1 7 7 0 42 93 94 0 13 44 0 0 1 1 0 4 14 5 (a) (b) Fig 2: Attribute Usage Values and Application Access Frequencies Let Q = {91;:::;44} be a set of queries, A = {A1; : : : ;A4} be a set of attributes, and S = {S1,S2,S3,S4) be a set of sites. The matrix of Figure 3(a) describes the attribute usage values and the matrix of Figure 3(b) gives the application access frequencies. Now, a) Calculated attribute affinity matrix for the above scenario. b) Suppose attribute A1, A2 is already placed in clustered affinity matrix. Now place A3 in the clustered affinity matrix. 51 52 S3 54 , , . , 1 0 1 0 91 41 9 11 0 10 e 72 0 1 1 0 8 12 19 0 0 1 0 1 7 7 0 42 93 94 0 13 44 0 0 1 1 0 4 14 5 (a) (b) Fig 2: Attribute Usage Values and Application Access Frequencies