J.L. Hammond, P. J. P. O'Reilly, "Performance Analysis of Local Computer Networks", Addison-Wesley Publishing Company, 1988.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....transmitted at rate C (bit sec) onto shared channel. We observed number of bits that are lost. In gure 3.3 we shows P loss info [bit] as a function of number of streams, J. We consider a scenario where: The theoretically maximal throughput which can be reached with unslotted ALOHA is 18 [7]. However it was under assumption that the arrival is a poisson process. In that case the probability of good transmission is 0.36, in other words, the loss probability is arround 0.64. As expected we also have big losses in our simulation because of the collisions, so the smoother will not help ....

....alpha2, error1 analysis, alpha3, error2 ideal csma Figure 4.11: Version II analysis, ploss vs. dj, a = 0.05 Copyright at Technical University Berlin. All Rights reserved. 4.2. 4 Classical CSMA CD analysis, version III Another variation is to follow the classical CSMA CD analysis as in [7]. They suggest that for every TRANSP of useful transmission, an overhead of h PROP is incurred, where h = 3 or 5 in literature. This gives : e E = h PROP h a L j C j (1 h a) 1 h a) 1 h a) 1e 07 1e 06 1e 05 0.0001 0.001 0.01 0.1 1 csma cd analysis, h = ....

J. L. Hammond and P. J. P. O'Reilly. Performance Analysis of Local Computer Networks. Addison{Wesley, 1986.

....Tx,Low TB,Low TB,Low 234 Priority 3. Delay Analysis In this section we determine the average transfer delay on the network measured from the time the packet is completely stored in Tx to the time the packet is completely received by Rx. We use the virtual cut through approach as described in [3] with similar assumptions: # The latency for each node is constant # The capacity of all buffers is infinite # The interarrival times and service times are independent (thus we can use independent submodels for each node) The following assumptions apply for high and low priority traffic ....

.... priority # in a priority queue with four (# ##highest priority) 2] ### # # ### # ### # # # (1) with the mean residual service time of # # (2) The average number ####### of TBs a packet experiences is for high and low priority traffic ####### # ## # ### ######## (3) From [3] this gives us a relation between the traffic in a TB and the traffic in a Tx: # ######### # ####### # ######### The network throughput of the buffer insertion ring has portions for high and low priority traffic # ### # ### ##### . Note that # ### accounts only for the part dedicated to one ....

J. L. Hammond and P. J. O'Reilly. Performance Analysis of Local Computer Networks. Addison-Wesley, 1988.

....from the OLT to every ONU and every ONU to the OLT. That is true for all PON topologies. Therefore, the OLT is the only device that can arbitrate the time divided access to the shared channel. In Section 3, we propose an OLT based polling scheme, similar to hub polling (see, for example [12]) Simple hub polling, however, has a disadvantage of increasing the polling cycle due to accumulation of walk times (switchover times) In access networks, this problem is even more pronounced than in LANs, as the propagation delays in the access network are much higher. For example, in a PON ....

J. L. Hammond and P. J. P. O'Reilly. Performance Analysis of Local Computer Networks, Addison Wesley, 1987.

....is usually given in bytes per second although it can also be in packets per second. Transfer delay is defined as the time from the arrival of the last bit of the packet into the station of a network until the last bit of this packet is delivered through the network to its destination station [4]. In a CSMA CD type of network, factors that contribute to the transfer delay are defer time (T D ) contention time, propagation delay ( P ) and transmission time (T X ) Adding all these factors will result to the transfer delay of a packet. Collision count is the number of collisions ....

Hammond, J.L. and P. J. P. O Reilly, Performance Analysis of Local Computer Networks , Addison-Wesley Publishing Company, Inc., 1986.

.... GammaaG ) 2 Gamma e GammaaG Gamma aGe GammaaG ) Figure 4 depicts the load throughput characteristics for the CSMA type schemes. Further reading: Numerous variations on the environment under which the Aloha and CSMA schemes operate have been addressed in the literature, see e.g. 3] [22], 23] 39] 46] For instance, various packet length distributions were considered in [2] 17] for Aloha and in [47] for CSMA. The assumption that whenever two or more packets overlap at the receiver, all packets are lost, is overly pessimistic. In radio networks the receiver might correctly ....

J.L. Hammond and P.J.P. O'Reilly, Performance Analysis of Local Computer Networks, Addison-Wesley Publishing Company, 1986.

....h) E [M ] h =N) 2) 4. Cambridge Ring The Cambridge ring [2] is an N station slotted ring with n slots. Each slot (or minipacket) is 38 bits in length and conveys m = 2 characters of data. We round minipacket length to ks = 5 characters (40 bits) to ignore all overhead (such as gap digits [18]) outside the n slots, making round trip time = ksn character times. We assume the N stations are evenly spaced with station to station delay of =N . With SR a source station marks its returning busy minipacket empty and releases it for another s use. Packets are fragmented into nG minipackets of ....

J. L. Hammond and P. J. P. O'Reilly, Performance Analysis of Local Computer Networks. Reading, MA: Addison--Wesley Publ. Co., 1986.

....0.29067145 0.26158 10 0.8 0.04021934 0.36440306 0.3306 10 0.9 0.03135553 0.49716024 0.45636 10 0.99 0.01149814 0.82571810 0. 71355 It is worth noting the differences between the IEEE 802.11 MAC protocol (basically a CSMA protocol) and a CSMA MAC protocol with the collision detection mechanism [Ham 88] In the former case, for a given M, by increasing the packet length (i.e. the q value) we obtain a decrease in the optimal p value. This is because, in this case, increasing q causes an increase in the collision part of the t v ; hence the balance point is obtained by reducing the collision ....

J.L. Hammond, P. J. P. O'Reilly, "Performance Analysis of Local Computer Networks", Addison-Wesley Publishing Company, 1988.

....Token Passing. This, in turn, depends heavily on the number of stations in the network, their generation (sampling) rate, the packet distribution, and the actual network bandwidth. 131 All of these parameters and their effects have been extensively studied and queueing models have been developed [Hammond,86 , Hayes,84, Takagi,88 ] So instead of using discrete event simulation, we will use results from these published models as black boxes for generating transmission delays for variable loads as would be generated by each configuration, see figure 5.6. These variable loads are obtained by assuming ....

Hammond, J. L. and O'Reilly, P. J., Performance Analysis of Local Computer Networks, Addison-Wesley, (1986).

....is assumed in accessing the temporal slots in w iso c channel. Taking expectations of the above mentioned times, we have: where: the average waiting time in the queue Q OST4 at the ordinary station can be calculated with the formula for a low priority queue in a HOL non preemptive M D 1 model [HO86], BG87] E T iso [ E W QOS iso [ E W AOS iso [ 2 T TR iso 2 E T D iso [ E W QMSR iso [ T MS iso E W QMST iso [ E W SYNC iso [ 5.15 ( E W QOS iso [ 2l i iso m ctr iso ( 2 2 1 r i iso ( 1 2r i iso ( l i iso m ....

....temporal slots in w iso c channel, i.e. every station is allowed to transmit its cells in idle slots as long as the number of transmitted cells in one frame does not exceed a fixed, preliminary decided number. E[W LM iso ] is calculated with the formula for a M D 1 model with limited service [HO86], RS90] and E[W BS iso ] is calculated with a formula which is similar to (5.16) The parameter n represents the number of stations in network; and L iso is the average arrival rate, at master station, for call setup (or disengagement) requests, which is given by r i iso l i iso ctr ....

[Article contains additional citation context not shown here]

Hammond, J.L. and O'Reilly, P.J.P., Performance Analysis of Local Computer Networks, Addison-Wesley Publishing Company, USA, 1986.

....handling policies at each node traversed, and on the network congestion. Thus, it is almost impossible to determine an exact analytical model of an Internet communication. An accepted approximate model consists of a network of queues, one for each node along the path connecting two computers (Hammond and O Reilly, 1988). According to this model, queuing and dynamic routing introduce the variable transmission delay, and network congestion may result in packets losses. Thus, the approximate Internet model is characterized by the statistics of the packet delay and of the packet losses. Values of these parameters ....

....Host 2 in Figure 5. However, if the number of nodes is sufficiently large, the time delay follows the Gaussian distribution of Figure 6, as it is expected in a network of independent 11 queues such as, for example, a connection between the University of Padova and the Jet Propulsion Laboratory (Hammond and O Reilly, 1988). Long distance connections imply larger packet delays and larger variations of the delay, since data packets must access crowded back bone segments. Furthermore, Table 1 shows that higher packet losses are to be expected even for low data rates when the connection involves an intercontinental ....

Hammond, J. and O'Reilly, P. (1988). Performance Analysis of Local Computer Networks.

....packet size connections. For these reasons, we decided to verify the results reported in [3] by testing the performance of Internet for different connections. An accepted Internet model consists of a network of queues, each queue relative to a node along the dynamic path connecting the two users [10]. The effects of queuing and dynamic routing is observed by the users as a variable transmission delay or jitter. The presence of a congested link results in an unpredictable deferral of the packet transmission at the node before the congested link. Since all Internet packets have limited life ....

....0 20 40 60 80 100 120 140 160 180 200 Interval [ms] Figure 4: RTT distribution for Helios. time (this is known as [M,M,1] queue) As the number of queues increases with the number of nodes traversed, the total delay is equal to the sum of an increasing number of exponentially distributed variables [10]. As a final step of the characterization of an Internet connection, we analyze the packet losses behavior, and we show that the loss rate depends on the connection lenght and on the probing rate. This result could be expected, since higher probing rates result in increased network congestion and ....

J.L. Hammond and O'Reilly P.J.P. Performance Analysis of Local Computer Networks. AddisonWesley, 1986.

....a heterogeneous network of machines. It serves as a conceptual starting point in compiling for load balancing and communication in such an environment. We consider heterogeneity in both the processor and the network dimensions. We take a very different approach than the work on stochastic models [6], which are intended to model This workwas supported in part by an NSF Research Initiation Award and ARPA contract F19628 94 C 0057. the performance behavior of the whole system with possibly many jobs running at the same time in an unpredictable way, and are, therefore, more detailed and ....

Joseph L. Hammond and Peter J.P. O'Reilly. Performance analysis of local computernetworks.Addison-Wesley,1986.

....Commands play an important role in the inter operability of domestic appliances. Therefore, the performance characteristics of the control channel are of vital importance. It is well known from the computer network field that MAC protocols have a great impact on the performance of networks (Hammond and O Reilly 1986) (Tobagi 1980) In this paper we investigate the throughput of the MAC protocol of a twisted pair home network by means of a semi Markov model. The model considers the network under a high load situation. The protocol was originally described in (Tritton 1991) The paper is organized as follows. ....

Hammond, J.L. and P.J.R. O'Reilly. 1986. Performance Analysis of Local Computer Networks. Addison-Wesley, Reading Massachusetts.

....as follows. We assume light system load (queueing occurs for access to the control channel, but access to data channels is immediate. When a packet arrives at a node, it is put into that node s arrival queue. Access to the control channel is round robin TDMA, and can be modeled as an M D 1 queue [9] D c = P c NP c 2 N ae c P c 2(1 Gamma ae c ) where P c is the length (in bits) of the control packet and D c is the mean delay (in bit times) for access to and transmission on the control channel. The load offered to the control channel is ae c , which is defined to be aeWP c P where ....

J. L. Hammond and P. J. P. O'Reilly, Performance Analysis of Local Computer Networks. Addison Wesley, 1986.

No context found.

J.L. Hammond, P. J. P. O'Reilly, "Performance Analysis of Local Computer Networks", Addison-Wesley Publishing Company, 1988.

No context found.

J.L. Hammond and P.J.P. O'Reilly, Performance Analysis of Local Computer Networks (Addison-Wesley Publishing Company, 1988).

No context found.

J.L. Hammond, P. J. P. O'Reilly, "Performance Analysis of Local Computer Networks", Addison-Wesley Publishing Company, 1988.

No context found.

J. Hammond and P. O'Reilly. Performance Analysis of Local Computer Networks, volume I. Addison-Wesley Publishing Company, 1986.

No context found.

J. L. Hammond and P. J. P. O'Reilly, "Performance Analysis of Local Computer Networks," Addison-Wesley, 1988.

No context found.

J. L. Hammond and P. J.P. O'Reilly, Performance Analysis of Local Computer Networks, Addison-Wesley, 1988.

No context found.

J. L. Hammond and P.J.P. O'Reilly, Performance Analysis of Local Computer Networks. Reading, MA: Addison-Wesley Publishing Company, 1986.

No context found.

J.L. Hammond and P.J.P. O'Reilly, Performance Analysis of Local Computer Networks (Addison-Wesley, Reading, MA, 1986).

No context found.

J. Hammond and P. O'Reilly "Performance Analysis of Local Computer Networks", Adison-Wesley 1986.

No context found.

Hammond, J.L. and P.J.P. O'Reilly, The Performance Analysis of Local Computer Networks, Addison-Wesley Publishing Company, Reading MA, 1986.

No context found.

Hammond, Joseph and Peter O'Reilly, Performance Analysis of Local Computer Networks, Addison-Wesley, Reading, Mass., 1986.

First 50 documents

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC