### Table 2 shows upper bounds for the number of messages originated in one round by two clock synchronization algorithms: 1) m-ICV, and 2) a unistep ICV algorithm employing the TCID com- munication protocols of [27] and any of the convergence functions in [14, 16, 22, 28, 43]. These bounds are respectively derived in [6] and [27], under the assumption that (M) node-disjoint paths are used to convey clock values reliably, where (M) is the connectivity of the network. Note that the numbers of messages in arbitrary-topology networks should be regarded as absolute upper bounds for both m-ICV and TCID ICV. In regular NCCNs such as tori and hypercubes, fewer messages are used by both m-ICV and TCID ICV. In all cases, m-ICV uses orders of magnitude fewer messages than TCID ICV. For example, the TCID ICV algorithm uses about 1; 000 times the number of messages of 3-ICV in a 32,768-node hypercube.

1998

"... In PAGE 25: ... Table2 : Number of messages per round (upper bounds) Another interesting aspect of m-ICV is that it increases signi cantly the communication locality for clock synchronization messages in regular NCCNs. As an example, assume that 2-ICV is used to synchronize an N-node hypercube, in which TSPs are mapped to a 2(b(log2 N)=2c) 2(d(log2 N)=2e) array.... In PAGE 32: ...Figure 1: Synchronization of groups via multistep interactive convergence Figure 2: Synchronization of a 16-node system with 2-ICV (C = clock, A = a-clock) Table 1: Trade-o between fault tolerance and communication cost (m = 2, N = 256) Figure 3: The jth round of m-ICV Figure 4: Algorithm m-ICV pseudocode Table2 : Number of messages per round (upper bounds)... ..."

Cited by 3

### Table 4: Arbitrary logical topology over random network case. Capacity constraint: The values of Table 3 are repeated here to show that improvement concerning the connectivity constraint is not achieved by relaxing the capacity constraint. Connectivity constraint: The percentage of bad elements (which would cause disconnections in the higher level network protected groups) is very low for PIW compared to SPR-P and SPR-CC. There is also a strong reduction achieved by PIW concerning the average number of demands which cannot be restored by the failure of such a bad element.

"... In PAGE 14: ...6 for SPR-P). Table4 shows the measurements for the arbitrary logical topology over random network and for the con- nectivity constraint. The values for the capacity con- straints are the same as those of Table 3 and are re- peated here to show that the improvement concerning the connectivity constraint is not achieved by relax- ing the capacity constraint, but that both are achieved jointly.... ..."

### Table 4: Simulated vs. theoretical traffic for arbitrary tree topologies. We assume a normal distribution.

2007

"... In PAGE 8: ... Here we validate this assumption, assuming a normal distribution for f. Table4 shows the results. In Table 4(a), we set the standard deviation to 1 and varied f.... In PAGE 8: ... Table 4 shows the results. In Table4 (a), we set the standard deviation to 1 and varied f. The values of f are derived by maintaining the fixed de- fault network size while increasing h in unit increments, therefore obtaining increasingly sparse trees.... In PAGE 8: ... The values of f are derived by maintaining the fixed de- fault network size while increasing h in unit increments, therefore obtaining increasingly sparse trees. In Table4 (b), instead, we set f = 10 and varied the standard deviation from 1 to 5. Recall that a standard deviation of 5 means that the number of children is within the interval [5; 15] for 68% of the nodes, and within [0; 25] for 99.... ..."

Cited by 1

### Table 2 : Bandwidth needed for a subset of the compression algorithms. Considering an arbitrary network topology based on a client-server architecture , we classify the communication between a server and a client as either connection oriented or connectionless. Both classes follow a store-and-forward approach. In the latter, in contrary to the former , the data packets (assuming packet switching instead of message switching) try each time to nd their way to the destination (through several point to point links) without following a prede ned path (either a circuit switched one or a virtual one). In addition to these, there is also the possibility of achieving communication through a multiple access link (e.g., shared bus), or a series of such links. In a multiple access link nodes contend in a way speci ed by a protocol, to exclusively use the medium for the purpose of transmiting their packets. More 2

1997

"... In PAGE 2: ...involving the use of Wavelets, Fractals, Neural nets, e.t.c.) promise to deliver movies with much less bandwidth than that required by MPEG-2, the need for high bandwidth still remains if what we want is a high Quality VoD Service (QoS)(see Table2 ) provided to a large number... ..."

Cited by 4

### Table 3 shows a comparison between the different solutions proposed in literature or commercially available.

"... In PAGE 6: ...able 2: Main characteristics of several WLAN market segments..............................................15 Table3 Comparison between the different on-chip communication solutions.... In PAGE 22: ... Table3 Comparison between the different on-chip communication solutions. Topology From the point of view of the topology, the proposed solutions cover the whole spectrum from a simple bus, hierarchy of buses, fixed topology networks, hierarchical networks and arbitrary topologies.... ..."

### Table 12: Properties of ad hoc networking

"... In PAGE 10: ...able 11. Challenges in concurrent engineering.................................................................... 59 Table12 .... In PAGE 75: ... The union of these forms an arbitrary topology, which may change rapidly and unpredictably. Table12 shows the properties of ad hoc networking, adapted from (Thonet, 2004). 1.... ..."

### Table 1: Network Topologies

"... In PAGE 3: ... Network model Two different network topologies have been considered for our study (Figure 2). Table1 reports the number of nodes N, the number of unidirectional links L, the hop count averaged among all node pairs h and the link capacity C (Mb/s). The reason to have two different topologies is that the smaller 7nodes topology could be implemented both in the simulation study and in a testbed (see II.... ..."

### Table 5: Network Topologies

1995

"... In PAGE 5: ... We used LOOM to experimentally calculate the best network topol- ogy (number of hidden units). Table5 lists the best topology 1Scaling reduces the side effects of scale differences between param- eters. Linear, square root, logarithm, and general data transformation are... ..."

Cited by 4

### Table 1. Network topologies

"... In PAGE 7: ...of different partitionings. To compare the different algorithms, we used the topolo- gies listed in Table1 . The names of the topologies roughly follow the convention: type-nodes-links-otherparam.... ..."

### Table 1. Network Topology.

1997

Cited by 3