Social Locality in Data Storage - Data Storage for Social Networks: A Socially Aware Approach - page 14

Fig. 2.1 Random vs. socially
aware(threeservers,one
primary and one replica per
user node): replicas are
shown in dashed circles and
primaries in solid circles .
( a ) Social graph ( b ) Primary
partition ( c ) Random
replication ( d ) Socially-aware
replication

a

b

c

d

Tabl e 2. 1 Socially-aware vs. random replication: read cost of each user
Method/User 1 2 3 4 5 6 7 8 9 10 Total

Random 2 3 1 2 3 2 4 4 2 2 25

i ll - re3113212312 9

2.1 Perfect vs. Imperfect Social Locality

Ideally, the preservation of social locality in a storage system should be perfect , i.e.,
every user must have its data located on the same server with all of its neighbor
users. Consequently, any read request by a user asking for its data and that of
its neighbors can be fulfilled efficiently by visiting only one server. Perfect social
locality, however, is not always possible if the number of replicas for each user