A. Heybey, "The network simulator," Laboratory for Computer Science, Massachusetts Institute of Technology,September 1990.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....is underdamped, and (II) # 0 ## l goal , the rate convergence is overdamped, where # h goal and # l goal are functions of Q goal , p, q, #,and#. The transient state analysis aims at characterizing the system dynamics while the multicast tree bottleneck 8 We used the NetSim package [19] for the simulations, and set the simulation parameters exactly the same as those used the analysis for comparison purposes. path is still in progress converging from one to another equilibrium state. Denote the transient state initial rate gain by # 0 ,and the new bottleneck s target rate gain ....

....0 = 2 ms. To study the worst case, set # 0 = # min 4 = min i2f1;###;ng f# i g and e # = # max 4 = max i2f1;###;ng f# i g of a multicast VC (Virtual Circuit) with n paths, and assume e # = 267 cells ms. Fig. 4(c) plots N , obtained numerically by Eq. 22) and simulations by NetSim [19], vs. ## max,#min # for different q. N is found to increase stepwise monotonically with ## max ,#min #. This is expected since a large variation in RM cell RTT requires more transient cycles to converge to the new optimal equilibrium state. A smaller q results in a fewer number of transient ....

[Article contains additional citation context not shown here]

A. Heybey, "The network simulator," Laboratory for Computer Science, Massachusetts Institute of Technology,September 1990.

....of the target system. We believe that good approximations can be obtained by conducting careful studies of the complete target system. Survey of some existing simulators In order to reduce programming work, we examined three simulators that were available to us: 1) Network Simulator from MIT [Hey 89, Mar 88] 2) DeNet from University of Wisconsin Madison [Liv 89] and (3) COMNET II.5 from CACI Products Company [CACI] Routing Project: Initial Design May 18, 1990 4 Network Simulator Network Simulator is a discrete event simulator designed for simulating a network of messagepassing ....

....Network Simulator does not provide any components to describe routing, hierarchies and address resolution. The simulator is written in C and runs under UNIX and X windows. The source code is available. The documentation inside the code is good, whereas the external documentation is poor [Hey 89, Mar 88] DeNet DeNet (Discrete Event Network) is a general purpose simulator. It views the target system as a set of objects that communicate with each other. Objects are defined as instances of Discrete Event Modules (DEVM) DEVMs can be arbitrarily complex. Discrete event connectors (DEVC) ....

A. Heybey, "The Network Simulator", Laboratory of Computer Science, Massachusetts Institute of Technology, October 1989.

....by retrofitting the BSD implementation into the x kernel. Our implementation of Vegas was derived by modifying Reno. 2.1 Simulator Many of the results reported in this paper were measured on a network simulator. Even though several good simulators are available e.g. REAL [8] and Netsim [3] we decided to build our own simulator based on the x kernel. In this environment, actual x kernel protocol implementations run on a simulated network. Specifically, the simulator supports multiple hosts, each running a full protocol stack (TEST TCP IP ETH) and several abstract link behaviors ....

A. Heybey. The network simulator. Technical report, MIT, Sept. 1990.

....of TCP to another Finally, there are other protocols (everything is a protocol in the x kernel) that model other types of transfers; for example, the MEGTEST protocol models TCP bulk transfers. 2.2. 4 Related Work Some of best known freely available network simulators are REAL [19] Netsim [10] and the recently released ns from LBL s Network Research Group. None of these simulators support common implementations of protocols, that is, direct execution. 33 Instead there is code to simulate the major characteristics of the protocol. One problem with this approach is that it misses some of ....

A. Heybey. The network simulator. Technical report, MIT, Sept. 1990.

....Many successful TCP performance evaluations have been conducted via simulation (e.g. FF96, Hoe96] A large variety of simulators for modeling internetworking protocols exist, and are currently used by researchers. Some of these tools are OpNet [Tec] x sim [BP96] the Network Simulator [Hey90] REAL [Kes88] ns [MF95] as well as specialized simulators written by various researchers for their own use. Picking a simulator to use is sometimes complicated. We encourage researchers to investigate various simulators and choose the one best geared towards the research at hand. For instance, ....

A. Heybey. The Network Simulator. Technical report, MIT, September 1990.

....threshold, which is an estimation of the equilibrium operating point, i.e. a packet leaves the network as a sender puts a packet into the network, is key to the performance of the algorithm. In Figure 1, 4 we show a segment number versus time graph of a 1 Mbyte transfer in the Netsim simulator [6], running BSD Net 2 TCP implementation. All the simulator experiments in this paper use a segment size of 1024 bytes, and the maximum window size for the connection is 50 segments. For details about the simulation environment and how to read the graph, see the appendix. In this implementation, ....

....experiments over the real networks and give a short tutorial on how to read the sequence number versus time graphs. Finally, we evaluate the limitations of the simulations. A.1 The Simulations Environment A.1. 1 The Simulator The results presented in Section 2 are based on simulations in Netsim[6], a packet by packet, event driven simulator. The simulator is based on the model that a computer network consists of an interconnected set of communication and switching components. The switching component in the simulator simulates a Drop Tail gateway. When the queues in the switching components ....

A. Heybey. The network simulator. Technical report, MIT, September 1990.

....derived by retrofitting the BSD implementation into the kernel. Our implementation of Vegas was derived by modifying Reno. A. Simulator Many of the results reported in this paper were obtained from a network simulator. Even though several good simulators are available e.g. REAL [12] and Netsim [5] we decided to build our own simulator based on the kernel. In this environment, actual kernel protocol implementations run on a simulated network. Specifically, the simulator supports multiple hosts, each running a full protocol stack (TEST TCP IP ETH) and several abstract link behaviors ....

A. Heybey, "The network simulator," MIT, Technical Report, Sept. 1990.

....to debug and test the final implementation of a protocol, when it is easy to introduce subtle implementation bugs. Nevertheless, the difficulties of setting up a real testbed with the desired features has stimulated the development of a number of network simulators, such as REAL [1] Netsim [2, 3] and ns [4] The x kernel framework [5] has also been used for the implementation and testing of network protocols. Experiments on network protocols are usually aimed to determine protocols behaviour in complex networks made of many nodes, routers and links, with different queueing policies, ....

A.Heybey: "The network simulator", Technical Report, MIT, Sept.1990

....access to network environments that are not easily found in the production Internet, such as satellite links. Network simulators do not generate real network traffic, but rather model traffic and major network components internally. The strength of network simulators (such as REAL [Kes88] NetSim [Hey90] and LBL s ns [MF95] is that they allow study of complex network topologies that are difficult to create using real networks. In addition, simulators are not limited by the speed of the hardware that makes up a network. Therefore, simulators provide an environment for studying high speed ....

A. Heybey. The Network Simulator. Technical report, MIT, September 1990.

....the problem, some simplifications can be made, e.g. considering all phenomena of interest (queueing, delays, bandwidth constraints) confined to one or a few bottleneck links, and trying to run experiments or simulations on such topologies. Packet level simulators have been built to this purpose[6, 4, 9], and successfully used in research. As an alternative, experiments on real systems are performed using modified routers acting as a flakeway , and configured to introduce delay, losses and other perturbations to the traffic. The tool proposed in this paper acts much like a flakeway, except that ....

A.Heybey, "The network simulator", Technical Report, MIT, Sept.1990

....120 160 200 240 280 320 360 400 Sending KB S 60 120 180 240 300 360 420 480 540 600 Figure 1: TCP Reno trace examples. 2.1 Simulator Many of the results reported in this paper were obtained from a network simulator. Even though several good simulators are available e.g. REAL [12] and Netsim [5] we decided to build our own simulator based on the x kernel. In this environment, actual x kernel protocol implementations run on a simulated network. Specifically, the simulator supports multiple hosts, each running a full protocol stack (TEST TCP IP ETH) and several abstract link ....

A. Heybey. The network simulator. Technical report, MIT, Sept. 1990.

....the one implementation of TCP to another. Finally, there are other protocols (everything is a protocol in the x kernel) that model other types of transfers; for example, TCP bulk transfers transfers. 2. 4 Related Work Some of best known freely available network simulators are REAL [10] Netsim [5] and the recently released ns from LBL s Network Research Group 1 . Neither simulator supports common implementations of protocols, that is, direct execution. Instead there is code to simulate the major characteristics of the protocol. One problem with this approach is that it misses some of the ....

A. Heybey. The network simulator. Technical report, MIT, Sept. 1990.

....Such factors are often hard to model accurately in simulators, possibly resulting in inaccurate or unreliable results. Nevertheless, the difficulties of setting up a real network with the desired features has stimulated the development of a number of network simulators, such as REAL [11] Netsim [7, 8] and ns [12] The x kernel framework [9] has also been used for the implementation and testing of network protocols. In most cases, the aim of experiments on network protocols is to determine their behaviour in a complex network made of many nodes, routers and links, with different queueing ....

A.Heybey, "The network simulator", Technical Report, MIT, Sept.1990

....by retrofitting the BSD implementation into the x kernel. Our implementation of Vegas was derived by modifying Reno. 2.1 Simulator Many of the results reported in this paper were obtained from a network simulator. Even though several good simulators are available e.g. REAL [9] and Netsim [2] we decided to build our own simulator based on the x kernel. 1 We limit our discussion to Reno, which is both newer and better performing than Tahoe. Also note that in terms of the congestion related algorithms, Reno is roughly equivalent to the Berkeley Network Release 2 (BNR2) ....

A. Heybey. The network simulator. Technical report, MIT, Sept. 1990.

No context found.

A. Heybey. The network simulator. Technical report, MIT, 1990.

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