Page 143 - ITU Kaleidoscope 2016

ICTs for a Sustainable World

4. SYSTEM MODEL
Cache management in ICN may acquire the Hypergraph [35]
characteristics, where an association is obtained between
on-path routers and the original server. Our proposed CPCE
strategy follows graph theory, called Hypergraph [35–37].
Let a network with routers (V) and connections (E) be rep-
resented as a Hypergraph (H) [35–38], such as: H=(V,E),
where V={v 1 ,v 2 , ..., v n }, and E={e 1 ,e 2 , ..., e m }. Therefore,
the network relationship can be defined as R={r 1 ,r 2 , ..., r n },
where r i is the i th router and n is the number of total
routers. Similarly, the connections denoted by E is such that
E={e 1 ,e 2 , ..., e m }, where e j is the j th connection and m is
the number of total connections.
The GEANT topology is used for the validation of the CPCE Figure 2. GEANT topology
strategy using Hypergraph, [39]:
A topology T, containing R routers and E connections as

T = {R, E} (1) d r = {In − degree + Out − degree} (5)
with the objective as GEANT maintains the ICN formulation For example, in Figure 2 d(r 4 )= 6 and d(r 14 )= 6 for 6 inter-
as connected routers on the edge.
According to the definition of the graph, the maximum degree

T = {R, E} (2) of a network G is represented by ΔG.
To prove our model using GEANT topology, the maximum
as GEANT consists of 22 nodes (see Figure 2), therefore
router degree is ΔT =6.
22
Therefore, the overall router degree, which is defined as max-
R = (r i ), (3) imum cache capacity of ICN router is given by
i=1
n
where r represents the number of nodes (routers), and as
ΔT = (r i ). (6)
the number of connections (E) in GEANT topology is 38,
i=1
therefore
To know the idea of ICN routers and connections relationship,
38
assume the routers’ membership in T, as described in [35,39],
E = (e j ). (4)
if r i represents routers and C i the cache size, it implies that
j=1
each router r i can cache a content. Then the cache size of
Due to the Internet heterogeneity, each router has a con- router r i can be
nection pair e j , such as: [e 1 = {r 0 ,r 1 }, e 2 = {r 1 ,r 2 },
n
e 3 = {r 2 ,r 3 }, e 4 = {r 3 ,r 4 }, e 5 = {r 3 ,r 5 }, e 6 =
C size = (C i ). (7)
{r 5 ,r 6 }, e 7 = {r 6 ,r 7 }, e 8 = {r 7 ,r 8 }, e 9 = {r 8 ,r 9 },
e 10 = {r 9 ,r 10 }, e 11 = {r 10 ,r 11 }, e 12 = {r 11 ,r 12 }, e 13 = i=1
{r 4 ,r 12 }, e 14 = {r 12 ,r 9 }, e 15 = {r 12 ,r 5 }, e 16 = {r 5 ,r 13 }, This develops a network topology T with routers and connec-
e 17 = {r 5 ,r 14 }, e 18 = {r 13 ,r 14 }, e 19 = {r 14 ,r 15 }, tions as
e 20 = {r 14 ,r 16 }, e 21 = {r 16 ,r 17 }, e 22 = {r 17 ,r 1 },
e 23 = {r 17 ,r 15 }, e 24 = {r 17 ,r 18 }, e 25 = {r 18 ,r 14 },
T = {r 1 ,r 2 , ..., r n }, (8)
e 26 = {r 18 ,r 19 }, e 27 = {r 18 ,r 20 }, e 28 = {r 19 ,r 20 },
e 29 = {r 20 ,r 21 }, e 30 = {r 21 ,r 14 }, e 31 = {r 21 ,r 0 }, e 32 = where rC.
{r 21 ,r 4 }, e 33 = {r 4 ,r 18 }, e 34 = {r 4 ,r 7 }, e 35 = {r 4 ,r 15 },
e 36 = {r 15 ,r 12 }, e 37 = {r 15 ,r 1 }, e 38 = {r 1 ,r 3 }].
Table 1. Simulation Scenario.
Thus by the definition of Hypergraph, directed and undirected
Cache Size 1GB-10GB
connections are achieved. In Figure 2, r 0 and r 1 have a direct 8
connection, whereas r 0 and r 3 have an undirect connection, Catalog Size 10
Zipf probability (α) 0.7, 1.0
therefore, it is generalized in the GEANT order that r i ,r i +1=
direct, otherwise, the connection is undirect . Topology GEANT and DTelekom
Social Network Topology Facebook [40–42]
In the case of ICN, a router represents the overall connectiv-
Simulator SocialCCNSim [43,44]
ity, i.e., intersection and inter-connectivity. Hence a router’s
Simulation Runs 10 times
degree can be represented as [39]:
– 125 –