Social Locality in Data Storage - Data Storage for Social Networks: A Socially Aware Approach - page 17

2.3 Optimization Objectives

Minimizing the server load and maximizing load balancing are among the most
important objectives of any distributed storage system. The server load is catego-
rized into three types: read load, write load, and storage load. We formulate these
objectives below.

2.3.1 Read Load

To understand the server read load , consider a read request for user i . This request
needs to be directed to its primary server, say server s, which will provide the data
for i . Suppose that user i is also interested in the data of user j , one of its neighbors.
To retrieve this data, there are two cases:

￿Userj 's data (primary or replica) is located on server s: The data of j can be

provided by server s, requiring no additional server request.
￿Userj 's data (primary or replica) is not located on server s: The data of j needs
to be retrieved from the primary server of j , requiring one additional read request
sent to this server.

The amount of data returned to user i is the same in both cases, but the number of
read requests that need to be processed at the server side is different, and, especially,
worse if i and j do not colocate. In OSNs where data under request are of small size,
having more read requests incurs more traffic and CPU processing at the server side.
Therefore, an important objective is to minimize the server load due to read requests,
which we refer to as read load , given the social relationships between the users and
the rate at which they initiate read requests.
Given a server s, its read load is computed as

0 1

X

N

X

N

X

M

read
s

@ p is C .1 p is / A :

l D r i p js e ij p it .1 x jt / (2.3)

i D 1 j D 1 t D 1

This load consists of read requests that are initiated by (1) users i primarily assigned
to s and (2) users i not primarily assigned to s, that have neighbors j primarily
assigned to s but these neighbors are not colocated with i . The number of read
requests belonging to the latter group depends on the social strength matrix E which
determines whether a neighbor's data needs also to be retrieved.
The total read load of all the servers is

0 1

X

M

X

M

X

N

X

N

X

M

@ p is C .1 p is / A

read D read

s

L l D r i p js e ij p it .1 x jt /

s D 1 s D 1 i D 1 j D 1 t D 1

(2.4)