### Table 1. Simulated clock speed for 1-D topology. Each row has fixed number of nodes per host and each column has fixed number of hosts. All value is normalized to real time clock speed.

2005

"... In PAGE 7: ... It constitutes the best possible case for the distributed simulation and as such pro- vides a rough upper bound on the performance. Table1 presents the results. Each cell of the table shows the ratio of the simulated average clock speed to the real time clock speed, of 7372800 cycles per second.... In PAGE 7: ... This relationship is predicted by the analysis of Theorem 2 presented in the previous section but none the less, we found the degree to which it holds somewhat surprising. By way of comparison to previous work, in this best case sce- nario 2048 nodes can be simulated at nearly a tenth of the real time speed using 16 hosts (lower righthand corner of Table1 ), which is almost 8 times better than results reported for TOSSIM [12]. Also, nearly 160 nodes can be simulated in real time speed using 16 hosts, and improvement of almost a factor of 5 over previous... In PAGE 8: ...f 2, 4, ..., and 2048 nodes respectively. The units of the y-axis on the lefthand side of the graph are for the ratio shown in Table1 . For each point, we also plot the corresponding host number at which the best performance is achieved (the host count is shown on the y- axis at the righthand side of the graph).... ..."

Cited by 3

### TABLE II SIMULATED CLOCK SPEED FOR 1 -D TOPOLOGY. EACH ROW HAS FIXED NUMBER OF NODES PER HOST AND EACH COLUMN HAS FIXED NUMBER OF HOSTS. ALL VALUE IS NORMALIZED TO REAL TIME CLOCK SPEED. Nodes Hosts number

2005

Cited by 3

### TABLE III SIMULATED CLOCK SPEED FOR 2 -D TOPOLOGY. EACH ROW HAS FIXED NUMBER OF NODES PER HOST AND EACH COLUMN HAS FIXED NUMBER OF HOSTS. ALL VALUE IS NORMALIZED TO REAL TIME CLOCK SPEED. Nodes Hosts number

2005

Cited by 3

### Table 3. Simulated clock speed for all-to-all complete graph. Each row has fixed number of nodes per host and each column has fixed number of hosts. All value is normalized to real time clock speed.

2005

"... In PAGE 8: ... For the worst case, we simulate an all-to-all complete graph configura- tion in which each simulated node must consider all of the other nodes to be in radio range making communication overhead maxi- mal. Table3 and Figure 8 shows the speedup factors and scalability curves respectively. In this worst case, communication overhead in-... ..."

Cited by 3

### TABLE IV SIMULATED CLOCK SPEED FOR ALL-TO-ALL COMPLETE GRAPH. EACH ROW HAS FIXED NUMBER OF NODES PER HOST AND EACH COLUMN HAS FIXED NUMBER OF HOSTS. ALL VALUE IS NORMALIZED TO REAL TIME CLOCK SPEED.

2005

Cited by 3

### TABLE VI SIMULATED CLOCK SPEED FOR SIMULATED 2 -D GRID OF MICAZ MOTES. EACH ROW HAS FIXED NUMBER OF NODES PER HOST AND EACH COLUMN HAS FIXED NUMBER OF HOSTS. ALL VALUES ARE NORMALIZED TO REAL TIME CLOCK SPEED.

2005

Cited by 3

### Table 4. Simulated clock speed for simulation of 2-D grid of Mica2 motes on CLUSTER2. Each row has fixed number of nodes per host and each column has fixed number of hosts. All values are normalized to real time clock speed.

2005

"... In PAGE 9: ... To test our simulator in a larger scale, we perform the 2-D ex- periment on CLUSTER2, a 64-node cluster. Table4 presents the results. Comparing Table 4 to Table 2 (which used CLUSTER1 for the same configuration) CLUSTER2 achieves lower speedup fac- tors for the test cases they have in common (columns 1 through 5).... In PAGE 9: ... Table 4 presents the results. Comparing Table4 to Table 2 (which used CLUSTER1 for the same configuration) CLUSTER2 achieves lower speedup fac- tors for the test cases they have in common (columns 1 through 5). Both a slower processor speed in CLUSTER2 and, somewhat curiously, higher network latency (even though CLUSTER1 and CLUSTER2 both use gigabit Ethernet as an interconnect) con- tribute to this lower performance.... ..."

Cited by 3

### TABLE V SIMULATED CLOCK SPEED FOR SIMULATION OF 2 -D GRID OF MICA2 MOTES ON CLUSTER2. EACH ROW HAS FIXED NUMBER OF NODES PER HOST AND EACH COLUMN HAS FIXED NUMBER OF HOSTS. ALL VALUES ARE NORMALIZED TO REAL TIME CLOCK SPEED. Nodes Hosts number

2005

Cited by 3

### Table 1. Comparison of VirtualClock and Fair Queueing algorithms (Ratio of real-time to best-effort traffic is 80:20.)

"... In PAGE 4: ... We mea- sured the inter-frame delivery time and standard deviation (SD) of delivery time for the media streams. The results with different input loads are quite similar in both cases as depicted in Table1 . However, since implementation of the Fair Queueing is more complex for maintaining the round robin number, we use the VirtualClock algorithm in the rest... ..."

### Table 1. Comparison of VirtualClock and Fair Queueing algorithms (Ratio of real-time to best-effort traffic is 80:20.)

"... In PAGE 4: ... We mea- sured the inter-frame delivery time and standard deviation (SD) of delivery time for the media streams. The results with different input loads are quite similar in both cases as depicted in Table1 . However, since implementation of the Fair Queueing is more complex for maintaining the round robin number, we use the VirtualClock algorithm in the rest... ..."