Results 11 - 20 of 4,244

Table 5.2.1 shows various performance parameters of the BHT per secure multicast session, where:

in MARKS: Zero Side Effect Multicast Key Management using Arbitrarily Revealed Key Sequences
by Ip Re England

Table 1, Number of packets and bytes needed for a number of different control tasks using unicast versus multicast in a session with 100 agents.

in Net-based Learning and the mStar Environment
by Kåre Synnes
"... In PAGE 113: ... In none of the tests lossy transcoding was performed and to be able to repeat the exact traffic, sessions were played back using the mMOD system [10]. Table1 presents how much bandwidth can be saved by using tunnel compression on the different data flows. The table shows that for normal data flows a compression ratio of about 5-9% is achieved.... In PAGE 113: ...0363% 0.0920% Table1... In PAGE 114: ... This section examines how much extra delay is actually introduced by the tunnel. The introduced delay was measured on the global lecture (row 4 in Table1 ). The measurements were done by measuring the round-trip-time (RTT) of a data flow between two hosts on different networks, one sender and one echo-host which bounces the received traffic back to the sender.... In PAGE 129: ... 2.4 Bandwidth Usage Table1 presents the number of messages and bandwidth consumed for doing a number of control tasks. Column 1: The task being performed (see below); 2: Total number of packets transmitted; 3: Number of bytes consumed without counting the underlying transport protocol; 4: Number of bytes consumed including the transport protocol using the reference implementation (see Section 4).... ..."

TABLE I NUMBER OF PACKETS AND BYTES NEEDED FOR A NUMBER OF DIFFERENT CONTROL TASKS USING UNICAST VERSUS MULTICAST IN A SESSION WITH 100 AGENTS.

in A Framework for Management and Control of Distributed Applications using Agents and IP-multicast
by Peter Parnes, Kåre Synnes, Dick Schefström

Table 5(b) shows the number of TCP losses that occur in the network when there are no multicast sessions. We can see from Tables 5(d) and (e), that TCP drops increase signi cantly in the presence of multicast ows. However, given the number of TCP packets sent, this is a rather small di erence in the overall TCP drop ratio. We can see that we have many tradeo that can be made on with regard to the number of representatives, the drop rate and the throughput. Generally, a few representatives work quite well, with larger representative sets making only a modest improvement. While the drop rates might be considered pessimistic compared to typical TCP drop rates, the drops are computed are the probability of any multicast packet being dropped on any link. Consequently, multicast drop rates cannot be compared directly with TCP loss rates.

in A Multicast Congestion Control Mechanism for Reliable Multicast
by Dante DeLucia , Katia Obraczka 1997
"... In PAGE 8: ....2.1 Single Multicast Flow The purpose of this experiment is to show that our congestion control algorithm detects and makes use of available bandwidth. The average throughput values shown in Table5 (a) demonstrate that the multicast ow is able to achieve close to maximum utilization on the 16 Kbytes/sec bottleneck links. The table also shows that the congestion control mechanism manages to keep the average drop ratio to zero, except for one case, where the average drop ratio is 0.... In PAGE 9: ... As shown in Figure 3(d), the congestion control algorithm manages to still keep the number of packets drop low. In fact, Table5 (d) shows that for the runs with 3 representatives and 75% density, both the multicast and the TCP drop ratios were kept quite low, under 4 and 1%, respectively. In bulk data transfer operations, it may be desirable to trade low packet loss for higher throughput.... In PAGE 9: ... Finally, the average feedback ratio is the average ratio of feedback packets received at the source to the number of multicast data packets sent. The average rates given in Table5 (c) and Table 5(e) are per multicast ow. The aggregate throughput achieved by both multicast ows is twice the listed throughput.... ..."
Cited by 27

Table 5(b) shows the number of TCP losses that occur in the network when there are no multicast sessions. We can see from Tables 5(d) and (e), that TCP drops increase signi cantly in the presence of multicast ows. However, given the number of TCP packets sent, this is a rather small di erence in the overall TCP drop ratio. We can see that we have many tradeo that can be made on with regard to the number of representatives, the drop rate and the throughput. Generally, a few representatives work quite well, with larger representative sets making only a modest improvement. While the drop rates might be considered pessimistic compared to typical TCP drop rates, the drops are computed are the probability of any multicast packet being dropped on any link. Consequently, multicast drop rates cannot be compared directly with TCP loss rates.

in A Multicast Congestion Control Mechanism for Reliable Multicast
by Dante Delucia, Katia Obraczka 1997
"... In PAGE 8: ....2.1 Single Multicast Flow The purpose of this experiment is to show that our congestion control algorithm detects and makes use of available bandwidth. The average throughput values shown in Table5 (a) demonstrate that the multicast ow is able to achieve close to maximum utilization on the 16 Kbytes/sec bottleneck links. The table also shows that the congestion control mechanism manages to keep the average drop ratio to zero, except for one case, where the average drop ratio is 0.... In PAGE 9: ... As shown in Figure 3(d), the congestion control algorithm manages to still keep the number of packets drop low. In fact, Table5 (d) shows that for the runs with 3 representatives and 75% density, both the multicast and the TCP drop ratios were kept quite low, under 4 and 1%, respectively. In bulk data transfer operations, it may be desirable to trade low packet loss for higher throughput.... In PAGE 9: ... Finally, the average feedback ratio is the average ratio of feedback packets received at the source to the number of multicast data packets sent. The average rates given in Table5 (c) and Table 5(e) are per multicast ow. The aggregate throughput achieved by both multicast ows is twice the listed throughput.... ..."
Cited by 27

Table 1 - Description of the protocol messages

in unknown title
by unknown authors 2000
"... In PAGE 7: ... Four messages are defined in the protocol, namely, ADVERT, INSTALL, ERROR, UNINSTALL. The messages are shown in Table1 and described next. ADVERT is a message issued by an upstream node to a downstream node to advertise a multicast session and a set of wavelengths that can be used to carry data of that multicast session.... ..."
Cited by 3

Table 27. Multicasting: 1 to N Scenario

in unknown title
by unknown authors 2000
"... In PAGE 10: ...able 26. Multicasting: 1 to N Scenario Applications................................................................. 94 Table27 .... In PAGE 103: ... Note that the merging of the presentation layer and session layer functions in the application layer eliminates the overhead due to these two layers and the need to set up a connection. Table27 and 29 present the results of the analysis. In these tables the functions related to the presentation and session layers are added to application layer and those that are not required are eliminated.... ..."

Table 1: Mean number of packets/link, session time of end-to-end multicast tree, L=2, 3; pD=1-(1-p)L. L = 2 L = 3

in Shortest-hop Based Reliable Network Multicast
by K. L. Eddie Law, Daniel Siu

Table 3. A list of demand sessions used in this example.

in Allocation of splitting nodes in alloptical wavelength-routed networks
by Maher Ali, Jitender Deogun 2000
"... In PAGE 11: ... 20. A set of session demands of size 20 (shown in Table3 ) is to be routed in the network. The number of multicast nodes, K, is equal to 3.... ..."
Cited by 1

Table 1. Constants and variables used in this paper.

in Allocation of splitting nodes in alloptical wavelength-routed networks
by Maher Ali, Jitender Deogun 2000
"... In PAGE 4: ... In the following, we develop an Integer-Linear Program (ILP) for solving the SP-WRN problem. Table1 lists the constants and variables used in the formulation. The set of multicast sessions requested is repre- sented by Q, that is, Q [q i 1fcig, where ci si; Di denotes a multicast session from station si to all destinations in the set Di.... ..."
Cited by 1
Results 11 - 20 of 4,244