### TABLE I SUMMARY OF DIRECTIONAL MAC PROTOCOL PERFORMANCE.

### TABLE I SUMMARY OF DIRECTIONAL MAC PROTOCOL PERFORMANCE.

### Table 2: Estimates of branching probabilities for MPEG mac- roblock decoding.

"... In PAGE 4: ... 3 represent basic flow of control decisions taken during the decoding of each macroblock within a picture. Table2 shows estimates of branching probabili- ties for these decision points. These estimates were generated by running a software decoder on a collection of MPEG video traces.... In PAGE 4: ... These estimates were generated by running a software decoder on a collection of MPEG video traces. The first two columns in Table2 identify the type of picture (I, P, or B) and the percentage of occurrence of that particular type of pic- ture in the fixed sequence of pictures considered for our MPEG-2 decoder. The third column in the table gives the branching proba- bility for the first decision point in Fig.... In PAGE 4: ...ility for the first decision point in Fig. 3, i.e., the probability that a macroblock will be skipped within a P or a B picture (note that all macroblocks within an I pictures are intra-coded). The three last columns in Table2 give the probability that a given macroblock will be intra-coded, forward/backward coded, or bidirectionally coded, for I, P and B pictures. The performance of an MPEG-2 decoder is determined by the individual performance of five key modules: Variable Length De- coding (VLD), Inverse Quantization (IQ), Inverse DCT (IDCT), Pixel Interpolation (PI), and Pixel Add (PA) [7].... In PAGE 5: ... In the upper part of Fig. 4, we show the decoding time distri- butions for I, P, and B pictures for design Option 1, derived using the execution delays per macroblock (in # cycles) given in Table 3, the branching probabilities given in Table2 , and the previously mentioned model for the VLC+IQ block. Table 4 shows the corre- sponding average and worst case decoding times (in # cycles) for the three types of pictures, and also the worst case and average de- coding time considering all picture types (given on the last row of the table).... ..."

### Table 2: Estimates of branching probabilities for MPEG mac- roblock decoding.

"... In PAGE 4: ... 3 represent basic flow of control decisions taken during the decoding of each macroblock within a picture. Table2 shows estimates of branching probabili- ties for these decision points. These estimates were generated by running a software decoder on a collection of MPEG video traces.... In PAGE 4: ... These estimates were generated by running a software decoder on a collection of MPEG video traces. The first two columns in Table2 identify the type of picture (I, P, or B) and the percentage of occurrence of that particular type of pic- ture in the fixed sequence of pictures considered for our MPEG-2 decoder. The third column in the table gives the branching proba- bility for the first decision point in Fig.... In PAGE 4: ...ility for the first decision point in Fig. 3, i.e., the probability that a macroblock will be skipped within a P or a B picture (note that all macroblocks within an I pictures are intra-coded). The three last columns in Table2 give the probability that a given macroblock will be intra-coded, forward/backward coded, or bidirectionally coded, for I, P and B pictures. The performance of an MPEG-2 decoder is determined by the individual performance of five key modules: Variable Length De- coding (VLD), Inverse Quantization (IQ), Inverse DCT (IDCT), Pixel Interpolation (PI), and Pixel Add (PA) [7].... In PAGE 5: ... In the upper part of Fig. 4, we show the decoding time distri- butions for I, P, and B pictures for design Option 1, derived using the execution delays per macroblock (in # cycles) given in Table 3, the branching probabilities given in Table2 , and the previously mentioned model for the VLC+IQ block. Table 4 shows the corre- sponding average and worst case decoding times (in # cycles) for the three types of pictures, and also the worst case and average de- coding time considering all picture types (given on the last row of the table).... ..."

### Table I. We have used this distribution to design Sift, a MAC protocol for wireless sensors [8]. Essentially, the protocol is

2004

Cited by 16

### Table 1: Notation used in analysis of the protocols.

"... In PAGE 5: ... The models of the protocols enable prediction of the behavior due to changes in various param- eters such as the number of nodes, the number of channels, and the average packet generation rate. Table1 summarizes the notation used in this section to derive the performance models. The performance metrics of primary concern are throughput and packet delay.... In PAGE 6: ... A semi- markov model is more general and allows arbitrary state holding times. Define = 1 ? e? as the probability of at least one packet arrival in one slot where is the Poisson arrival rate per node; let Ps denote the probability of successful packet transmission and let Pb be the probability of leaving backoff state and attempting retransmission ( Table1 ). The transition probability matrix P is given below, where pij indicates the probability of transition from state Si to Sj: P = 0 B B B B @ 0 0 0 Ps Ps 0 (1 ? Ps) (1 ? Ps) 0 Ps 0 0 (1 ? Ps) 0 Pb Pb (1 ? Pb) (1 ? Pb) 0 0 Pb 0 (1 ? Pb) 1 C C C C A where = 1 ? .... ..."

### Table 1: Design activities identified in the protocol analysis

"... In PAGE 1: ...Table1... In PAGE 5: ... There is a mapping between knowledge of design activities and knowledge of knowledge t ransformers (A d :: K t) showing the inextricable link between the design and learning activities at the epistemic level. Depending on the type of the design activity ( Table1 ), there are two possible ways by which design and learning goals interact with o ne another; this depicts the inextricable link at the teleological level. Firstly, the learning goal (which is triggered by the rationale learning trigger, T lw ) precedes the design goal (i.... ..."

### Table 2: Comparision of MAC Protocols Based on No Pretransmission Coordination

"... In PAGE 4: ... The three supported classes are connection-oriented with guaran- teed bandwidth, connection-oriented without guaranteed bandwidth, and datagram traffic. 3 Protocols Based on No Pretransmission Coordination The protocols that do not require pretransmission coordination are summarized in Table2 . The tell and go column is not shown because it is not necessary for a station to tell the destination that it is transmitting.... ..."

### Table 4: Rejection ratio and disconnect ratio for different NAT/firewall solutions.

"... In PAGE 5: ... 5.2 Internet Experiments Table4 shows the rejection ratio for the different solutions. Each value is the average of three runs.... In PAGE 5: ...he rejection ratio from .08 to .02 in the Basic Contributor solution. However, Enhanced Contributor does well without any Connectivity- Aware Structuring so there is no room for it to help. Table4 also shows the average and 95 percentile disconnect ratio. The trend shown by this metric is similar to rejection rate, with Strawman having a... ..."