χ=5 virtual time state LVT entirely saved state partially saved state χ=5 ν=2 virtual time state LVT entirely saved partially saved unsaved state

Size: px
Start display at page:

Download "χ=5 virtual time state LVT entirely saved state partially saved state χ=5 ν=2 virtual time state LVT entirely saved partially saved unsaved state"

Transcription

1 ROLLBACK-BASED PARALLEL DISCRETE EVENT SIMULATION BY USING HYBRID STATE SAVING Francesco Quaglia Dipartimento di Informatica e Sistemistica, Universita di Roma "La Sapienza" Via Salaria 113, Roma, Italy, quaglia@dis.uniroma1.it Vittorio Cortellessa Dipartimento di Informatica S&P, CERTIA Research Center, Universita di Roma Torvergata Via della Ricerca Scientica, Roma, Italy, cortelle@info.utovrm.it KEYWORDS Parallel discrete event simulation, hybrid saving, rollback-recovery mechanisms, performance evaluation. ABSTRACT Optimistically synchronized parallel discrete event simulators must sometimes undo, by rolling back parts of the system, the erroneous over optimistic computation deriving from the decentralized management of the event list. For this reason, an essential part of these simulators is the saving mechanism. Three saving mechanisms have been proposed in literature: copy, periodic and incremental saving. In this paper we introduce a new saving technique, that will be referred to as hybrid, which mixes the advantages of previous approaches. We also present experimental results obtained in a simulation environment which adopts hybrid saving; such results quantify the benets, in terms of reduced simulation execution time, achievable by using our technique. 1 INTRODUCTION In parallel discrete event simulation the simulation program is partitioned into a number of logical processes (LPs), which model the behavior of dierent parts of the simulated system (Fujimoto 1990). The interaction between LPs is realized by message exchange; messages are stamped with a virtual time value (timestamp) that indicates when, in virtual time, the receiving LP must process them. The processing of a message determines the execution of an event which moves an LP from one to another and eventually produces some new messages that can This work is partially supported by Scientic Cooperation Network of the European Community \OLOS" under contract No. ERB4050PL and by the University of Roma at "Tor Vergata" CERTIA Research Center Project on Multimedia and Collaborative Technology, the MURST Projects on Performability in Software Engineering and Performance of Client-Server Systems, and the CNR Project on Performance and Reliability Engineering of Distributed Databases be addressed to any LP. Each LP has its own simulation clock (local virtual time), and an event list in which incoming messages are enqueued. Optimistic approach allows each LP to execute asynchronously. However, in order to ensure correct simulation results, certain causality constraints must be met. Specically, each LP must process received messages in non decreasing timestamp order. The most common optimistic method is referred to as Time Warp (Jeerson 1985). Under Time Warp no LP undergoes constraints in order to process a message, thus local simulation clocks may diverge. For this reason an LP may receive a message with a timestamp that is smaller than its current local virtual time (straggler message), breaking the causality constraints between events. In this case, the over-optimistic portion of the simulation is undone by rolling back the LP to a previous. From this the LP resumes its computation. The implementation of the rollback procedure requires the ability of restoring, at run time, a past of an LP. To this purpose, a simple mechanism, often called copy saving, has been proposed (Jefferson and Sowizral 1982; Jeerson 1985). It consists of copying into a queue the entire of an LP each time it executes an event. According to this solution the to be restored (due to rollback) is always available, but the saving overhead usually reaches unacceptable levels. Commonly, rollbacks are unfrequent compared to ordinary event execution, so, given that an LP can be regenerated from an earlier one by simply reexecuting some intermediate events, the periodic saving technique has been introduced for reducing the overhead due to copy saving (Bellenot 1992; Fujimoto 1990). According to this solution only a subset of LP s are saved, and the number of events executed between successive saving operations is known as checkpoint interval of an LP. For simulation models in which at each event execution only a fraction of the variables of the LP are modied, a third saving technique has been proposed, namely incremental saving (Steinman 1993; Unger et al. 1993). This solution consists of saving, at each event execution, only the inverse of

2 the changes of an LP, so that a past can be regenerated by applying, one at a time, the saved changes in backward order. In this paper we propose a new saving technique, that will be referred to as hybrid saving, which mixes the benecial eects of both periodic and incremental saving. According to this technique, an LP periodically saves its, but it has also the ability of saving the inverse of the incremental changes of its occurring in a checkpoint interval. In this way an LP is able to regenerate a past either by starting from an earlier saved and re-executing some intermediate events, or by applying to a later saved its previous inverse changes. We compare our technique to the above mentioned ones which, as it will be shown, can be obtained as a particular case of hybrid saving. We both design and propose an implementation of the hybrid saving protocol and we show experimental results which quantify the advantages, in terms of reduced simulation execution time, achievable by using such protocol compared to preexisting proposals. The paper is organized as follows: in Section 2 we describe actually used saving techniques, in Section 3 we introduce the hybrid saving protocol and we present a simulation environment adopting hybrid saving, experimental results which quantify the performance improvements achievable by using our technique are shown in Section 4, short conclusions constitute Section 5. 2 BACKGROUND In this section we give some details of actually used saving mechanisms in Time Warp simulators. 2.1 Copy and Periodic State Saving When copy saving is adopted, the of an LP is saved into a queue before a new event is executed. Usually, the saved is also marked with the local virtual time () of the LP (that is the timestamp of the last processed message). In this way all the s passed through by an LP are available and the rolling back of the LP to a virtual time T is realized by simply restoring the most recent marked with smaller than T. Upon rolling back, the LP cancels from the queue all the recorded s with larger than or equal to T. This technique adds to each event execution a checkpointing overhead, which is quantied by the sum of the time required to allocate a buer and the time required to copy the current into the buer. One approach for reducing such overhead is to perform saving every event executions ( being the checkpoint interval of an LP). This solution, known as periodic saving, has both benecial and detrimental eects on the LP execution time. It reduces the number of CPU cycles spent in saving operations, but upon rolling back, the required may not be in the queue. In latter case, such must be recomputed by reprocessing input messages (thus adding a time penalty). An LP that is recomputing a missing is said to be in a coasting forward phase (Fujimoto 1990). Both analytical and experimental studies have been carried out (Lin et al. 1993; Palaniswamy and Wilsey 1993a; Preiss et al. 1994) to point out the relation between periodic saving and the simulation execution time of an LP; several techniques for allowing an LP to dynamically recalculate the value for its checkpoint interval have also been proposed (Fleischmann and Wilsey 1995; Palaniswamy and Wilsey 1993b; Quaglia and Auriche 1997; Ronngren and Ayani 1994), in order to reduce the simulation execution time compared to the one obtained with static periodic sate saving. 2.2 Incremental State Saving Many challenging simulations (for example simulations of large communication systems) are characterized by LPs with large event execution time and with very large (up to hundreds of kilobytes), where only fractions of the are updated at each event execution. In these simulations it may be very inef- cient both saving copies of the complete, and regenerating a past by starting from an earlier one (through reprocessing intermediate events). In such simulations it is often protable to use incremental saving, in which only the inverse of the changes are saved at each event execution. Thus the saving mechanism builds a chain of inverse of the changes that, in case of rollback, allows the reconstruction of a past by starting from the current of the LP and applying the saved changes in backward order till the required is obtained. This solution successfully reduces the saving and reconstruction overhead when both the number of inverse changes saved at each event execution and the rollback length are minimal (Palaniswamy and Wilsey 1993a). 3 HYBRID STATE SAVING As shown in the previous section, an LP running in a Time Warp simulation must be able to regenerate whichever past, in order to ensure the correctness of the simulation rollback. It can be achieved either by starting from an older saved and reprocessing intermediate events or by backward applying saved incremental changes to the current. The mixing of such mechanisms consists in allowing an LP to start from whichever saved and to regenerate the required either by reprocessing

3 Τ1 Τ entirely saved χ=5 Τ2 partially saved virtual time Figure 1: an example of hybrid saving Τ1 unsaved Τ χ=5 entirely saved ν=2 Τ2 partially saved virtual time Figure 2: hybrid saving with = 5 and = 2 events (forward regeneration) or by backward applying saved incremental changes (backward regeneration). This mixed approach can be achieved by periodically saving the LP, and by saving the inverse of the changes in each checkpoint interval. In Figure 1 we show a portion of the evolution of an LP (i.e., a portion of the s passed through by the LP) which records whole its each ve event executions ( = 5), and saves all the inverse of the changes that occur in a checkpoint interval. In the rest of the paper we denote as partially saved the copy of the variables which are going to be modied by the next event execution, while we denote as entirely saved the copy of the whole LP. The LP whose evolution is shown in Figure 1 is able to regenerate the with LV T = T either by starting from the with LV T = T 1 and reprocessing two events, or by starting from the with LV T = T 2 and applying three inverse changes to the variables. However, by saving all the inverse of the incremental changes that occur in a checkpoint interval, only the length of the LP checkpoint interval can be tuned for reducing the total overhead due to the rollback-recovery mechanism. A generalization of the structure shown in Figure 1 is easily obtained by allowing an LP to save inverse changes of its starting from whichever in the checkpoint interval. In Figure 2 we show a case in which the LP saves its entire after ve event executions, while it partially saves two s in a checkpoint interval. In this case two parameters can be tuned: the checkpoint interval and the number of partially saved in each checkpoint interval. On the other hand, in this case the LP is not able to regenerate the with LV T = T by starting from the with LV T = T 2 and applying inverse changes, but it must necessarily regenerate that by starting from the with LV T = T 1 and by executing a coasting forward phase. Given that, at most,?1 s are partially saved in each checkpoint interval we get: 0 = < 1. As last consideration, copy, periodic and incremental saving can be generated starting from hybrid saving and by assigning appropriate values to the parameters and. In particular: - when = 1 (thus is necessarily 0) we get copy saving; - when > 1 and = 0 we get periodic saving; - when =? 1 and goes to innity we get incremental saving. We have implemented hybrid saving on the distributed simulation platform SIMCOR (Ciciani and Angelaccio 1994). This platform has been realized for carrying out parallel optimistically synchronized simulations on the hypercubical machine ipsc/2 (the platform's software is realized in C). In SIMCOR, LPs are statically assigned to processors; the aggressive approach (Gafni 1985) is adopted in the cancellation phase (i.e., antimessages are sent as soon as an LP rolls back). A single scheduler (Time Warp kernel) runs on each processor and manages the local event list by scheduling local LPs according to the STF (Shortest-Timestamp-First) algorithm (Jefferson 1985). The scheduler also manages the queue of each LP. In this way, both the scheduling and the saving mechanisms are transparent to the LPs. Inter-process communication is realized by using routines which are supported by the NX/2 distributed operating system running on the ipsc/2 machine. NX/2 also supports synchronization primitives which are used for computing GVT and executing fossil collection of obsolete totally/partially saved s and buered messages. In SIMCOR each message is stamped with its own type: for example, in queueing networks, the NEW- CUSTOMER type schedules the arrival of a new customer, while the END-SERVICE type schedules the end of the service for a customer. 3.1 The State Saving Protocol On each processor, the scheduler manages two arrays: the i-th entry of one array records the checkpoint interval of the i-th LP; the i-th entry of the other array records the number of the inverse of the changes that must be saved in a checkpoint interval for the i-th LP. In addition, each LP is endowed with a counter denoting the number of processed events from the last entirely saving operation of the LP. Each time the scheduler extracts a new event for a local LP (and before the LP executes such event), it checks (by using the corresponding counter) whether

4 pointer to the previous buffer of the LP F new buffer pointer to the new buffer Figure 3: structure of a buer the whole of the LP must be saved (i.e., if events have been processed from the last entirely saving ), or just the inverse of the changes must be recorded (i.e., if more than?? 1 events have been processed from the last entirely saving ), or no saving operation must be executed. In the rst case, the scheduler dynamically allocates a new buer in the LP queue and copies the current process into the buer. The buer is also marked with the of the scheduled LP and with a ag F which indicates whether the buer contains a copy of the entire or not (the structure of the buer is shown in Figure 3). In the second case, all updates to variables are identied and a new buer is allocated and linked to the queue only for back up copies of the variables which will be modied by the event execution. In SIMCOR, the identication of the variables that will be modied results quite simple because each message is stamped with its own type; the scheduler, by checking the type of the extracted message, identies the portion of code that will be executed by the LP, thus the corresponding part of the that will be modied. All the buers (containing either an entire or parts of the ) are linked by back pointers. 3.2 The State Regeneration Protocol By adopting hybrid saving a process may regenerate a past either by coasting forward some intermediate events, or by backward applying to a saved its previous inverse changes. Although the rst choice is always feasible, the second one works only if the required past falls in the range covered by the inverse changes saved in a checkpoint interval (this problem has already been shown above). In SIMCOR the protocol for regenerating a past is deterministic and works as follows: if the required is out of the ones covered by starting from a saved and backward applying inverse changes, normal coasting forward is used; otherwise, the is regenerated by backward reconstruction. Furthermore, in case of rollback, the regeneration is executed in atomic fashion (message/antimessage preemption is discarded). If the LP must rollback to the virtual time T, the scheduler searches, into the queue, the couple (S',S") of successive entirely saved such that LV T S 0 < T < LV T S 00 (note that the S" may be not yet saved into the queue; in this case the current LP is considered instead of S"). Then the oldest buer, if any, containing inverse of the changes between S' and S" is searched (S ), and its is compared with T. If T < LV T S then a coasting forward phase is required; the scheduler restores the S' in the LP and forces it to replay some already processed messages. Otherwise the scheduler starts to apply inverse incremental changes to the S" by backward running through the LP queue. Such procedure stops when the rst buer marked with LV T < T is found. The obtained is then restored in the LP. After the LP rolls back, all the buers with larger than T are released. 4 PERFORMANCE EVALUA- TION In this section we propose several experiments to show the benets, in terms of reduced execution time, that can be achieved by using hybrid saving. We compare the simulation execution time spent when hybrid saving is adopted to the one spent by using periodic saving. We propose simulation results of a stochastic queueing model, whose topology is a fully connected net with 32 service centers, where a constant customer population circulates among the centers (three customers in each center at the simulation starting). The timestamp increments are taken from an exponential distribution with mean 1 unit time, and customers are equally likely to be forwarded to any other center. We denote as s the time to save the entire copy of the LP and with e the average time to execute one event (excluding the time for sending the eventually produced messages). In the simulated model, only an average portion of 1=3 of the is modied by the execution of an event, so we approximatively have a s =3 average time spent for saving the inverse of a change (both the saving of the entire and the saving of a portion of the need the dynamic allocation of the corresponding buer, so the time for saving a fraction of the is not exactly proportional to the fraction of the to be saved). We focused our attention on the execution time required for committed events. The proposed set of experiments has been realized by adopting two dierent values of the ratio s = e : s = e = 2 and s = e = 1=2 (the value of the ratio s = e is modied, as in (Preiss et al. 1994), by introducing a variable delay loop into the event execution routine). In this way we can point out an idea of the performance of hybrid saving either when the saving cost dominates or when the event execution cost dominates. We studied the execution time of both periodic saving (referred to as P SS) and hybrid saving (referred to as HSS) while varying the checkpoint interval of the LPs. In the case of HSS, we plot two curves: HSS 1, where = b=2c, and HSS 2, where =? 1.

5 In Figure 4 ( s = e = 2) and in Figure 5 ( s = e = 1=2), the execution times obtained with P SS, HSS 1 and HSS 2 are shown vs. the length of the checkpoint interval of the LPs. The execution time results as the average of 20 runs, and the measures in dierent runs were within 5% of each other of some experiments carried out in a simulation environment supporting hybrid saving. Future work could be focused on both dening the convenient application domain of hybrid saving (e.g., while varying the portion of the modied by the execution of one event), and building algorithms for the dynamic selection of the protocol parameters ( and ). execution time (sec) PSS HSS1 HSS checkpoint interval of the LPs Figure 4: execution time when s = e = 2 The results show that, for this simulation model, HSS improves performances especially for large values of the checkpoint interval. This is because, when is large, P SS suers from an high overhead due to the coasting forward, while HSS reduces such overhead especially for large values of (in fact, the plots show that HSS 2 performs better than HSS 1 ). This phenomenon is clearly evident when considering large event execution time, so the gain achievable by using hybrid saving is larger in the case of s = e = 1=2. In conclusion, the results point out that when grows, the advantages introduced by the reconstruction mechanism of HSS, makes the overhead due to rollback-recovery ever smaller than the one of P SS. 5 CONCLUSIONS In this paper we have introduced a new saving protocol (namely hybrid) for Time Warp simulators. This protocol merges the advantages of the most commonly adopted saving techniques (periodic and incremental). Hybrid saving gains over other techniques by reducing the overhead time due to the rollback-recovery mechanism. The performance improvements achievable by adopting our approach are quantied by the results execution time (sec) PSS HSS1 HSS checkpoint interval of the LPs Figure 5: execution time when s = e = 1=2 REFERENCES Bellenot, S "State skipping performance with the Time Warp operating system", In Proceedings of 1992 SCS Workshop on Parallel and Distributed Simulation (Newport Beach, California, January 20-22). Society for Computer Simulation, Ciciani, B. and M. Angelaccio "An interface to develop Time-Warp based parallel simulations". In Proceedings of 1994 Massively Parallel Processing Conference (Delft, Holland, June 21-23), Elsevier Science, Fleischmann, J. and P.A. Wilsey "Comparative analysis of periodic saving techniques in Time Warp simulators". In Proceedings of 1995 SCS Workshop on Parallel and Distributed Simulation (Lake Placid, New York, June 14-16). Society for Computer Simulation, Fujimoto, R.M "Parallel discrete event simulation", Communications of ACM 33, no.10 (October): Gafni, A "Space management and cancellation mechanisms for Time Warp", Tech. Rep. TR University of Southern California, Los Angeles, California. Jeerson, D. and H. Sowizral "Fast concurrent simulation using the Time Warp mechanism; part I: local control", Tech. Rep. N1906AF. RAND Corporation, (December). Jeerson, D "Virtual time", ACM Trans. on Programming Languages and Systems 7, no.3 (July): Lin, Y.B.; B.R. Preiss; W.M. Loucks and E.D. Lazowska "Selecting the checkpoint interval in Time Warp simulation". In Proceedings of 1993 SCS Workshop on Parallel and Distributed Simulation (San Diego, California, May 17-19). Society for Computer Simulation, Palaniswamy, A.C. and P.A. Wilsey. 1993a. "An analytical comparison of periodic checkpointing and incremental saving". In Proceedings of 1993 SCS Workshop on Parallel and Distributed Simulation (San Diego, California, May 17-19). Society for Computer Simulation, Palaniswamy, A.C. and P.A. Wilsey. 1993b. "Adaptive checkpoint intervals in an optimistically synchronized parallel digital system simulator". In Proceedings of IFIP TC/WG10.5 Int. Conf. on Very Large Scale Integration (September) Preiss, B.R.; W.M. Loucks and D. MacIntyre "Eect of the checkpoint interval on time and space in Time Warp". ACM Transactions on Modeling and Computer Simulation 4, no.3 (July): pp Quaglia, F. and L.R.G. Auriche "A new technique for adaptive checkpointing in Time Warp". In Proceedings of 1997 SCS European Simulation Multiconference (Istanbul, Turkey, June 1-4). Society for Computer Simulation, Ronngren, R. and R. Ayani "Adaptive checkpointing in Time Warp". In Proceedings of 1994 SCS Workshop on Parallel and Distributed Simulation (Edinburgh, Scotland, July 6-8). Society for Computer Simulation, Steinman, J "Incremental saving in SPEEDS using C plus plus". In Proceedings of 1993 Winter Simulation Conference (Los Angeles, California, December). Society for Computer Simulation, Unger, B.W.; J.G. Cleary; A. Covington and D. West "External management system for optimistic parallel simulation". In Proceedings of 1993 Winter Simulation Conference (Los Angeles, California, December). Society for Computer Simulation,

Rollback Overhead Reduction Methods for Time Warp Distributed Simulation

Rollback Overhead Reduction Methods for Time Warp Distributed Simulation Rollback Overhead Reduction Methods for Time Warp Distributed Simulation M.S. Balsamo and C. Manconi* Dipartimento di Matematica e Informatica, University of Udine Vial delle Scienze 108, Udine, Italy,

More information

This article appeared in Proc. 7th IEEE Symposium on Computers and Communications, Taormina/Giardini Naxos, Italy, July , IEEE Computer

This article appeared in Proc. 7th IEEE Symposium on Computers and Communications, Taormina/Giardini Naxos, Italy, July , IEEE Computer This article appeared in Proc. 7th IEEE Symposium on Computers and Communications, Taormina/Giardini Naxos, Italy, July 1-4 2002, IEEE Computer Society. Software Supports for Preemptive Rollback in Optimistic

More information

An Empirical Performance Study of Connection Oriented Time Warp Parallel Simulation

An Empirical Performance Study of Connection Oriented Time Warp Parallel Simulation 230 The International Arab Journal of Information Technology, Vol. 6, No. 3, July 2009 An Empirical Performance Study of Connection Oriented Time Warp Parallel Simulation Ali Al-Humaimidi and Hussam Ramadan

More information

AGGRESSIVENESS/RISK EFFECTS BASED SCHEDULING IN TIME WARP

AGGRESSIVENESS/RISK EFFECTS BASED SCHEDULING IN TIME WARP Proceedings of the 2000 Winter Simulation Conference J. A. Joines, R. R. Barton, K. Kang, and P. A. Fishwick, eds. AGGRESSIVENESS/RISK EFFECTS BASED SCHEDULING IN TIME WARP Vittorio Cortellessa Computer

More information

Eect of fan-out on the Performance of a. Single-message cancellation scheme. Atul Prakash (Contact Author) Gwo-baw Wu. Seema Jetli

Eect of fan-out on the Performance of a. Single-message cancellation scheme. Atul Prakash (Contact Author) Gwo-baw Wu. Seema Jetli Eect of fan-out on the Performance of a Single-message cancellation scheme Atul Prakash (Contact Author) Gwo-baw Wu Seema Jetli Department of Electrical Engineering and Computer Science University of Michigan,

More information

Event List Management In Distributed Simulation

Event List Management In Distributed Simulation Event List Management In Distributed Simulation Jörgen Dahl ½, Malolan Chetlur ¾, and Philip A Wilsey ½ ½ Experimental Computing Laboratory, Dept of ECECS, PO Box 20030, Cincinnati, OH 522 0030, philipwilsey@ieeeorg

More information

Comparative Analysis of Periodic State Saving Techniques in Time. Warp Simulators. Center for Digital Systems Engineering. Cincinnati, Ohio

Comparative Analysis of Periodic State Saving Techniques in Time. Warp Simulators. Center for Digital Systems Engineering. Cincinnati, Ohio This paper appeared in the Proceedings of the 9th Workshop on Parallel and Distributed Simulation, PADS-1995. c 1995, IEEE. Personal use of this material is permitted. However, permission to reprint or

More information

Optimistic Distributed Simulation Based on Transitive Dependency. Tracking. Dept. of Computer Sci. AT&T Labs-Research Dept. of Elect. & Comp.

Optimistic Distributed Simulation Based on Transitive Dependency. Tracking. Dept. of Computer Sci. AT&T Labs-Research Dept. of Elect. & Comp. Optimistic Distributed Simulation Based on Transitive Dependency Tracking Om P. Damani Yi-Min Wang Vijay K. Garg Dept. of Computer Sci. AT&T Labs-Research Dept. of Elect. & Comp. Eng Uni. of Texas at Austin

More information

JWarp: a Java library for parallel discrete-event simulations

JWarp: a Java library for parallel discrete-event simulations CONCURRENCY: PRACTICE AND EXPERIENCE Concurrency: Pract. Exper.,Vol.10(11 13), 999 1005 (1998) JWarp: a Java library for parallel discrete-event simulations PEDRO BIZARRO,LUÍS M. SILVA AND JOÃO GABRIEL

More information

Consistent Logical Checkpointing. Nitin H. Vaidya. Texas A&M University. Phone: Fax:

Consistent Logical Checkpointing. Nitin H. Vaidya. Texas A&M University. Phone: Fax: Consistent Logical Checkpointing Nitin H. Vaidya Department of Computer Science Texas A&M University College Station, TX 77843-3112 hone: 409-845-0512 Fax: 409-847-8578 E-mail: vaidya@cs.tamu.edu Technical

More information

TIME WARP PARALLEL LOGIC SIMULATION ON A DISTRIBUTED MEMORY MULTIPROCESSOR. Peter Luksch, Holger Weitlich

TIME WARP PARALLEL LOGIC SIMULATION ON A DISTRIBUTED MEMORY MULTIPROCESSOR. Peter Luksch, Holger Weitlich TIME WARP PARALLEL LOGIC SIMULATION ON A DISTRIBUTED MEMORY MULTIPROCESSOR ABSTRACT Peter Luksch, Holger Weitlich Department of Computer Science, Munich University of Technology P.O. Box, D-W-8-Munchen,

More information

Transparently Mixing Undo Logs and Software Reversibility for State Recovery in Optimistic PDES

Transparently Mixing Undo Logs and Software Reversibility for State Recovery in Optimistic PDES 0 Transparently Mixing Undo Logs and Software Reversibility for State Recovery in Optimistic PDES DAVIDE CINGOLANI, Sapienza University of Rome ALESSANDRO PELLEGRINI, Sapienza University of Rome FRANCESCO

More information

Other Optimistic Mechanisms, Memory Management!

Other Optimistic Mechanisms, Memory Management! Other Optimistic Mechanisms, Memory Management! Richard M. Fujimoto! Professor!! Computational Science and Engineering Division! College of Computing! Georgia Institute of Technology! Atlanta, GA 30332-0765,

More information

Just-In-Time Cloning

Just-In-Time Cloning Just-In-Time Cloning Maria Hybinette Computer Science Department University of Georgia Athens, GA 30602-7404, USA maria@cs.uga.edu Abstract In this work we focus on a new technique for making cloning of

More information

Event Reconstruction in Time Warp

Event Reconstruction in Time Warp Event Reconstruction in Time Warp Lijun Li and Carl Tropper School of Computer Science McGill University Montreal, Canada lli22, carl@cs.mcgill.ca Abstract In optimistic simulations, checkpointing techniques

More information

On Checkpoint Latency. Nitin H. Vaidya. In the past, a large number of researchers have analyzed. the checkpointing and rollback recovery scheme

On Checkpoint Latency. Nitin H. Vaidya. In the past, a large number of researchers have analyzed. the checkpointing and rollback recovery scheme On Checkpoint Latency Nitin H. Vaidya Department of Computer Science Texas A&M University College Station, TX 77843-3112 E-mail: vaidya@cs.tamu.edu Web: http://www.cs.tamu.edu/faculty/vaidya/ Abstract

More information

Chair for Network Architectures and Services Prof. Carle Department of Computer Science TU München. Parallel simulation

Chair for Network Architectures and Services Prof. Carle Department of Computer Science TU München. Parallel simulation Chair for Network Architectures and Services Prof. Carle Department of Computer Science TU München Parallel simulation Most slides/figures borrowed from Richard Fujimoto Parallel simulation: Summary/Outline

More information

The Impact of Lookahead on the Performance of Conservative Distributed Simulation

The Impact of Lookahead on the Performance of Conservative Distributed Simulation The Impact of Lookahead on the Performance of Conservative Distributed Simulation Bruno R Preiss Wayne M Loucks Department of Electrical and Computer Engineering University of Waterloo, Waterloo, Ontario,

More information

Lookahead Accumulation in Conservative Parallel Discrete Event Simulation.

Lookahead Accumulation in Conservative Parallel Discrete Event Simulation. Lookahead Accumulation in Conservative Parallel Discrete Event Simulation. Jan Lemeire, Wouter Brissinck, Erik Dirkx Parallel Systems lab, Vrije Universiteit Brussel (VUB) Brussels, Belgium {jlemeire,

More information

Optimistic Parallel Simulation of TCP/IP over ATM networks

Optimistic Parallel Simulation of TCP/IP over ATM networks Optimistic Parallel Simulation of TCP/IP over ATM networks M.S. Oral Examination November 1, 2000 Ming Chong mchang@ittc.ukans.edu 1 Introduction parallel simulation ProTEuS Agenda Georgia Tech. Time Warp

More information

Automatic Incremental State Saving

Automatic Incremental State Saving Automatic Incremental State Saving Darrin West Science Applications International Corporation 4301 N Fairfax Drive Arlington VA 22203 west@jade.std.saic.com Abstract We present an Incremental State Saving

More information

global checkpoint and recovery line interchangeably). When processes take local checkpoint independently, a rollback might force the computation to it

global checkpoint and recovery line interchangeably). When processes take local checkpoint independently, a rollback might force the computation to it Checkpointing Protocols in Distributed Systems with Mobile Hosts: a Performance Analysis F. Quaglia, B. Ciciani, R. Baldoni Dipartimento di Informatica e Sistemistica Universita di Roma "La Sapienza" Via

More information

Parallel Discrete Event Simulation

Parallel Discrete Event Simulation Parallel Discrete Event Simulation Dr.N.Sairam & Dr.R.Seethalakshmi School of Computing, SASTRA Univeristy, Thanjavur-613401. Joint Initiative of IITs and IISc Funded by MHRD Page 1 of 8 Contents 1. Parallel

More information

USING GENETIC ALGORITHMS TO LIMIT THE OPTIMISM IN TIME WARP. Jun Wang Carl Tropper

USING GENETIC ALGORITHMS TO LIMIT THE OPTIMISM IN TIME WARP. Jun Wang Carl Tropper Proceedings of the 2009 Winter Simulation Conference M. D. Rossetti, R. R. Hill, B. Johansson, A. Dunkin, and R. G. Ingalls, eds. USING GENETIC ALGORITHMS TO LIMIT THE OPTIMISM IN TIME WARP Jun Wang Carl

More information

Ecient Redo Processing in. Jun-Lin Lin. Xi Li. Southern Methodist University

Ecient Redo Processing in. Jun-Lin Lin. Xi Li. Southern Methodist University Technical Report 96-CSE-13 Ecient Redo Processing in Main Memory Databases by Jun-Lin Lin Margaret H. Dunham Xi Li Department of Computer Science and Engineering Southern Methodist University Dallas, Texas

More information

Computing Global Virtual Time in Shared- Memory Multiprocessors

Computing Global Virtual Time in Shared- Memory Multiprocessors Computing Global Virtual Time in Shared- Memory Multiprocessors RICHARD M. FUJIMOTO and MARIA HYBINETTE Georgia Institute of Technology Global virtual time (GVT) is used in the Time Warp synchronization

More information

MANUFACTURING SIMULATION USING BSP TIME WARP WITH VARIABLE NUMBERS OF PROCESSORS

MANUFACTURING SIMULATION USING BSP TIME WARP WITH VARIABLE NUMBERS OF PROCESSORS MANUFACTURING SIMULATION USING BSP TIME WARP WITH VARIABLE NUMBERS OF PROCESSORS Malcolm Yoke Hean Low Programming Research Group, Computing Laboratory, University of Oxford Wolfson Building, Parks Road,

More information

Real-Time Scalability of Nested Spin Locks. Hiroaki Takada and Ken Sakamura. Faculty of Science, University of Tokyo

Real-Time Scalability of Nested Spin Locks. Hiroaki Takada and Ken Sakamura. Faculty of Science, University of Tokyo Real-Time Scalability of Nested Spin Locks Hiroaki Takada and Ken Sakamura Department of Information Science, Faculty of Science, University of Tokyo 7-3-1, Hongo, Bunkyo-ku, Tokyo 113, Japan Abstract

More information

An Approach for Federating Parallel Simulators

An Approach for Federating Parallel Simulators An Approach for Federating Parallel Simulators Steve L. Ferenci Kalyan S. Perumalla Richard M. Fujimoto College Of Computing Georgia Institute of Technology Atlanta, GA 30332-0280 {ferenci,kalyan,fujimoto}@cc.gatech.edu

More information

Sapienza University of Rome

Sapienza University of Rome Sapienza University of Rome Ph.D. program in Computer Engineering XXV Cycle - 2012/3 Design of Software Support Structures for High Performance Optimistic Simulations with Special Focus on Multi-Core Hosting

More information

s00(0) s10 (0) (0,0) (0,1) (0,1) (7) s20(0)

s00(0) s10 (0) (0,0) (0,1) (0,1) (7) s20(0) Fault-Tolerant Distributed Simulation Om. P. Damani Dept. of Computer Sciences Vijay K.Garg Dept. of Elec. and Comp. Eng. University of Texas at Austin, Austin, TX, 78712 http://maple.ece.utexas.edu/ Abstract

More information

LANGUAGE BASED STATE SAVING EXTENSIONS FOR OPTIMISTIC PARALLEL SIMULATION

LANGUAGE BASED STATE SAVING EXTENSIONS FOR OPTIMISTIC PARALLEL SIMULATION Proceedings of the 1996 Winter S1:rnulation C 1 0nfereTlce ed. J. lvi. Charnes, D. J. l\lorrice, D. T. Brunner, and J. J. Snrain LANGUAGE BASED STATE SAVING EXTENSIONS FOR OPTIMISTIC PARALLEL SIMULATION

More information

PARALLEL SIMULATION. Rassul Ayani Department of Teleinformatics, Computer Systems Laboratory Royal Institute of Technology (KTH) Stockholm, Sweden

PARALLEL SIMULATION. Rassul Ayani Department of Teleinformatics, Computer Systems Laboratory Royal Institute of Technology (KTH) Stockholm, Sweden PARALLEL SIMULATION Rassul Ayani Department of Teleinformatics, Computer Systems Laboratory Royal Institute of Technology (KTH) Stockholm, Sweden Abstract This tutorial surveys various approaches to executing

More information

Computing Global Virtual Time!

Computing Global Virtual Time! Computing Global Virtual Time Issues and Some Solutions Richard M. Fujimoto Professor Computational Science and Engineering Division College of Computing Georgia Institute of Technology Atlanta, GA 30332-0765,

More information

Distributed Recovery with K-Optimistic Logging. Yi-Min Wang Om P. Damani Vijay K. Garg

Distributed Recovery with K-Optimistic Logging. Yi-Min Wang Om P. Damani Vijay K. Garg Distributed Recovery with K-Optimistic Logging Yi-Min Wang Om P. Damani Vijay K. Garg Abstract Fault-tolerance techniques based on checkpointing and message logging have been increasingly used in real-world

More information

Fault-Tolerant Computer Systems ECE 60872/CS Recovery

Fault-Tolerant Computer Systems ECE 60872/CS Recovery Fault-Tolerant Computer Systems ECE 60872/CS 59000 Recovery Saurabh Bagchi School of Electrical & Computer Engineering Purdue University Slides based on ECE442 at the University of Illinois taught by Profs.

More information

COMPILED CODE IN DISTRIBUTED LOGIC SIMULATION. Jun Wang Carl Tropper. School of Computer Science McGill University Montreal, Quebec, CANADA H3A2A6

COMPILED CODE IN DISTRIBUTED LOGIC SIMULATION. Jun Wang Carl Tropper. School of Computer Science McGill University Montreal, Quebec, CANADA H3A2A6 Proceedings of the 2006 Winter Simulation Conference L. F. Perrone, F. P. Wieland, J. Liu, B. G. Lawson, D. M. Nicol, and R. M. Fujimoto, eds. COMPILED CODE IN DISTRIBUTED LOGIC SIMULATION Jun Wang Carl

More information

CHAPTER 4 AN INTEGRATED APPROACH OF PERFORMANCE PREDICTION ON NETWORKS OF WORKSTATIONS. Xiaodong Zhang and Yongsheng Song

CHAPTER 4 AN INTEGRATED APPROACH OF PERFORMANCE PREDICTION ON NETWORKS OF WORKSTATIONS. Xiaodong Zhang and Yongsheng Song CHAPTER 4 AN INTEGRATED APPROACH OF PERFORMANCE PREDICTION ON NETWORKS OF WORKSTATIONS Xiaodong Zhang and Yongsheng Song 1. INTRODUCTION Networks of Workstations (NOW) have become important distributed

More information

Consistent Checkpointing in Distributed Computations: Theoretical Results and Protocols

Consistent Checkpointing in Distributed Computations: Theoretical Results and Protocols Università degli Studi di Roma La Sapienza Dottorato di Ricerca in Ingegneria Informatica XI Ciclo 1999 Consistent Checkpointing in Distributed Computations: Theoretical Results and Protocols Francesco

More information

Georgia Tech Time Warp (GTW Version 3.1) Programmer's Manual for Distributed Network of. Richard M. Fujimoto, Samir R. Das, Kiran S.

Georgia Tech Time Warp (GTW Version 3.1) Programmer's Manual for Distributed Network of. Richard M. Fujimoto, Samir R. Das, Kiran S. Georgia Tech Time Warp (GTW Version 3.1) Programmer's Manual for Distributed Network of Workstations Richard M. Fujimoto, Samir R. Das, Kiran S. Panesar, Maria Hybinette and Chris Carothers College of

More information

FAST CELL LEVEL ATM NETWORK SIMULATION

FAST CELL LEVEL ATM NETWORK SIMULATION Proceedings of the 22 Winter Simulation Conference E. Yücesan, C.-H. Chen, J. L. Snowdon, and J. M. Charnes, eds. FAST CELL LEVEL ATM NETWORK SIMULATION Xiao Zhong-e Rob Simmonds Brian Unger Dept. Computer

More information

DISTRIBUTED SELF-SIMULATION OF HOLONIC MANUFACTURING SYSTEMS

DISTRIBUTED SELF-SIMULATION OF HOLONIC MANUFACTURING SYSTEMS DISTRIBUTED SELF-SIMULATION OF HOLONIC MANUFACTURING SYSTEMS Naoki Imasaki I, Ambalavanar Tharumarajah 2, Shinsuke Tamura 3 J Toshiba Corporation, Japan, naoki.imasaki@toshiba.co.jp 2 CSIRO Manufacturing

More information

Event Ordering. Greg Bilodeau CS 5204 November 3, 2009

Event Ordering. Greg Bilodeau CS 5204 November 3, 2009 Greg Bilodeau CS 5204 November 3, 2009 Fault Tolerance How do we prepare for rollback and recovery in a distributed system? How do we ensure the proper processing order of communications between distributed

More information

Time-Event based processing, a Survey Bruce Jones Vanessa Wallace

Time-Event based processing, a Survey Bruce Jones Vanessa Wallace Time-Event based processing, a Survey Bruce Jones Vanessa Wallace 1. Introduction Event based processing is used in many computational processes, whether by intent or as a result of the task requirement.

More information

Parallel and Distributed VHDL Simulation

Parallel and Distributed VHDL Simulation Parallel and Distributed VHDL Simulation Dragos Lungeanu Deptartment of Computer Science University of Iowa C.J. chard Shi Department of Electrical Engineering University of Washington Abstract This paper

More information

Three Models. 1. Time Order 2. Distributed Algorithms 3. Nature of Distributed Systems1. DEPT. OF Comp Sc. and Engg., IIT Delhi

Three Models. 1. Time Order 2. Distributed Algorithms 3. Nature of Distributed Systems1. DEPT. OF Comp Sc. and Engg., IIT Delhi DEPT. OF Comp Sc. and Engg., IIT Delhi Three Models 1. CSV888 - Distributed Systems 1. Time Order 2. Distributed Algorithms 3. Nature of Distributed Systems1 Index - Models to study [2] 1. LAN based systems

More information

Checkpointing and Rollback Recovery in Distributed Systems: Existing Solutions, Open Issues and Proposed Solutions

Checkpointing and Rollback Recovery in Distributed Systems: Existing Solutions, Open Issues and Proposed Solutions Checkpointing and Rollback Recovery in Distributed Systems: Existing Solutions, Open Issues and Proposed Solutions D. Manivannan Department of Computer Science University of Kentucky Lexington, KY 40506

More information

Recovering from Main-Memory Lapses. H.V. Jagadish Avi Silberschatz S. Sudarshan. AT&T Bell Labs. 600 Mountain Ave., Murray Hill, NJ 07974

Recovering from Main-Memory Lapses. H.V. Jagadish Avi Silberschatz S. Sudarshan. AT&T Bell Labs. 600 Mountain Ave., Murray Hill, NJ 07974 Recovering from Main-Memory Lapses H.V. Jagadish Avi Silberschatz S. Sudarshan AT&T Bell Labs. 600 Mountain Ave., Murray Hill, NJ 07974 fjag,silber,sudarshag@allegra.att.com Abstract Recovery activities,

More information

FAULT TOLERANT SYSTEMS

FAULT TOLERANT SYSTEMS FAULT TOLERANT SYSTEMS http://www.ecs.umass.edu/ece/koren/faulttolerantsystems Part 16 - Checkpointing I Chapter 6 - Checkpointing Part.16.1 Failure During Program Execution Computers today are much faster,

More information

warped: A Time Warp Simulation Kernel for Analysis and Application Development Dale E. Martin, Timothy J. McBrayer, and Philip A.

warped: A Time Warp Simulation Kernel for Analysis and Application Development Dale E. Martin, Timothy J. McBrayer, and Philip A. Published in the Proceedings of the Hawaiian International Conference on System Sciences, HICSS-1996. c 1996, IEEE. Personal use of this material is permitted. However permission to reprint or republish

More information

Kevin Skadron. 18 April Abstract. higher rate of failure requires eective fault-tolerance. Asynchronous consistent checkpointing oers a

Kevin Skadron. 18 April Abstract. higher rate of failure requires eective fault-tolerance. Asynchronous consistent checkpointing oers a Asynchronous Checkpointing for PVM Requires Message-Logging Kevin Skadron 18 April 1994 Abstract Distributed computing using networked workstations oers cost-ecient parallel computing, but the higher rate

More information

A ROLLBACK MANAGER FOR OPTMISTIC DISTRIBUTED HLA SIMULATIONS

A ROLLBACK MANAGER FOR OPTMISTIC DISTRIBUTED HLA SIMULATIONS ua Ktraes ae computadores 183 A ROLLBACK MANAGER FOR OPTMISTIC DISTRIBUTED HLA SIMULATIONS Fernando Vardânega, Carlos Maziero Programa de PósGraduação em Informática Aplicada Pontifícia Universidade Católica

More information

Parallel Logic Simulation of VLSI Systems

Parallel Logic Simulation of VLSI Systems Parallel Logic Simulation of VLSI Systems Roger D. Chamberlain Computer and Communications Research Center Department of Electrical Engineering Washington University, St. Louis, Missouri Abstract Design

More information

Scalability of Parallel Simulation Cloning

Scalability of Parallel Simulation Cloning Scalability of Parallel Simulation Cloning Maria Hybinette Computer Science Department University of Georgia Athens, GA 30602-7404, USA maria@cs.uga.edu Richard M. Fujimoto College of Computing Georgia

More information

Speculative High-Performance Simulation. Alessandro Pellegrini A.Y. 2017/2018

Speculative High-Performance Simulation. Alessandro Pellegrini A.Y. 2017/2018 Speculative High-Performance Simulation Alessandro Pellegrini A.Y. 2017/2018 Simulation From latin simulare (to mimic or to fake) It is the imitation of a real-world process' or system's operation over

More information

FB(9,3) Figure 1(a). A 4-by-4 Benes network. Figure 1(b). An FB(4, 2) network. Figure 2. An FB(27, 3) network

FB(9,3) Figure 1(a). A 4-by-4 Benes network. Figure 1(b). An FB(4, 2) network. Figure 2. An FB(27, 3) network Congestion-free Routing of Streaming Multimedia Content in BMIN-based Parallel Systems Harish Sethu Department of Electrical and Computer Engineering Drexel University Philadelphia, PA 19104, USA sethu@ece.drexel.edu

More information

DISTRIBUTED SIMULATION SYSTEMS. Richard M. Fujimoto. College of Computing Georgia Institute of Technology Atlanta, GA 30332, U.S.A.

DISTRIBUTED SIMULATION SYSTEMS. Richard M. Fujimoto. College of Computing Georgia Institute of Technology Atlanta, GA 30332, U.S.A. Proceedings of the 2003 Winter Simulation Conference S. Chick, P. J. Sánchez, D. Ferrin, and D. J. Morrice, eds. DISTRIBUTED SIMULATION SYSTEMS Richard M. Fujimoto College of Computing Georgia Institute

More information

Lecture 2: September 9

Lecture 2: September 9 CMPSCI 377 Operating Systems Fall 2010 Lecture 2: September 9 Lecturer: Prashant Shenoy TA: Antony Partensky & Tim Wood 2.1 OS & Computer Architecture The operating system is the interface between a user

More information

PARALLEL AND DISTRIBUTED SIMULATION. Richard M. Fujimoto. College of Computing Georgia Institute of Technology Atlanta, GA 3033, U.S.A.

PARALLEL AND DISTRIBUTED SIMULATION. Richard M. Fujimoto. College of Computing Georgia Institute of Technology Atlanta, GA 3033, U.S.A. Proceedings of the 1999 Winter Simulation Conference P. A. Farrington, H. B. Nembhard, D. T. Sturrock, and G. W. Evans, eds. PARALLEL AND DISTRIBUTED SIMULATION Richard M. Fujimoto College of Computing

More information

Real-Time Coordination in Distributed Multimedia Systems

Real-Time Coordination in Distributed Multimedia Systems Real-Time Coordination in Distributed Multimedia Systems Theophilos A. Limniotes and George A. Papadopoulos Department of Computer Science University of Cyprus 75 Kallipoleos Str, P.O.B. 20537 CY-1678

More information

messages from disque to parsim messages from parsim to disque

messages from disque to parsim messages from parsim to disque Extension to DISQUE - A trace facility to produce trace data for use by a monitoring tool for distributed simulators Gerd Meister Department of Computer Science, University of Kaiserslautern P.O.Box 3049,

More information

Heckaton. SQL Server's Memory Optimized OLTP Engine

Heckaton. SQL Server's Memory Optimized OLTP Engine Heckaton SQL Server's Memory Optimized OLTP Engine Agenda Introduction to Hekaton Design Consideration High Level Architecture Storage and Indexing Query Processing Transaction Management Transaction Durability

More information

TRANSACTION-ORIENTED SIMULATION IN AD HOC GRIDS: DESIGN AND EXPERIENCE

TRANSACTION-ORIENTED SIMULATION IN AD HOC GRIDS: DESIGN AND EXPERIENCE TRANSACTION-ORIENTED SIMULATION IN AD HOC GRIDS: DESIGN AND EXPERIENCE Gerald Krafft and Vladimir Getov Harrow School of Computer Science University of Westminster Watford Rd, Northwick Park, Harrow HA1

More information

Performance Predictions for Speculative, Synchronous, VLSI Logic Simulation. Bradley L. Noble J. Cris Wade Roger D. Chamberlain

Performance Predictions for Speculative, Synchronous, VLSI Logic Simulation. Bradley L. Noble J. Cris Wade Roger D. Chamberlain Performance Predictions for Speculative, Synchronous, VLSI Logic Simulation Bradley L. Noble J. Cris Wade Roger D. Chamberlain Bradley L. Noble, J. Cris Wade, and Roger D. Chamberlain, Performance Predictions

More information

Performance Evaluation of Two New Disk Scheduling Algorithms. for Real-Time Systems. Department of Computer & Information Science

Performance Evaluation of Two New Disk Scheduling Algorithms. for Real-Time Systems. Department of Computer & Information Science Performance Evaluation of Two New Disk Scheduling Algorithms for Real-Time Systems Shenze Chen James F. Kurose John A. Stankovic Don Towsley Department of Computer & Information Science University of Massachusetts

More information

Telecommunication and Informatics University of North Carolina, Technical University of Gdansk Charlotte, NC 28223, USA

Telecommunication and Informatics University of North Carolina, Technical University of Gdansk Charlotte, NC 28223, USA A Decoder-based Evolutionary Algorithm for Constrained Parameter Optimization Problems S lawomir Kozie l 1 and Zbigniew Michalewicz 2 1 Department of Electronics, 2 Department of Computer Science, Telecommunication

More information

Redo Log Undo Log. Redo Log Undo Log. Redo Log Tail Volatile Store. Pers. Redo Log

Redo Log Undo Log. Redo Log Undo Log. Redo Log Tail Volatile Store. Pers. Redo Log Recovering from Main-Memory Lapses H.V. Jagadish AT&T Research Murray Hill, NJ 07974 jag@research.att.com Avi Silberschatz Bell Laboratories Murray Hill, NJ 07974 avi@bell-labs.com S. Sudarshan Indian

More information

An Efficient Log.Based Crash Recovery Scheme for Nested Transactions

An Efficient Log.Based Crash Recovery Scheme for Nested Transactions Microprocessing and Microprogramming 31 (1991) 99-104 99 North-Holland An Efficient Log.Based Crash Recovery Scheme for Nested Transactions Dong C. Shin and Song C. Moon Department of Computer Science

More information

Distributed Database Management System UNIT-2. Concurrency Control. Transaction ACID rules. MCA 325, Distributed DBMS And Object Oriented Databases

Distributed Database Management System UNIT-2. Concurrency Control. Transaction ACID rules. MCA 325, Distributed DBMS And Object Oriented Databases Distributed Database Management System UNIT-2 Bharati Vidyapeeth s Institute of Computer Applications and Management, New Delhi-63,By Shivendra Goel. U2.1 Concurrency Control Concurrency control is a method

More information

Some Thoughts on Distributed Recovery. (preliminary version) Nitin H. Vaidya. Texas A&M University. Phone:

Some Thoughts on Distributed Recovery. (preliminary version) Nitin H. Vaidya. Texas A&M University. Phone: Some Thoughts on Distributed Recovery (preliminary version) Nitin H. Vaidya Department of Computer Science Texas A&M University College Station, TX 77843-3112 Phone: 409-845-0512 Fax: 409-847-8578 E-mail:

More information

Design Patterns for Real-Time Computer Music Systems

Design Patterns for Real-Time Computer Music Systems Design Patterns for Real-Time Computer Music Systems Roger B. Dannenberg and Ross Bencina 4 September 2005 This document contains a set of design patterns for real time systems, particularly for computer

More information

Steering. Stream. User Interface. Stream. Manager. Interaction Managers. Snapshot. Stream

Steering. Stream. User Interface. Stream. Manager. Interaction Managers. Snapshot. Stream Agent Roles in Snapshot Assembly Delbert Hart Dept. of Computer Science Washington University in St. Louis St. Louis, MO 63130 hart@cs.wustl.edu Eileen Kraemer Dept. of Computer Science University of Georgia

More information

Availability of Coding Based Replication Schemes. Gagan Agrawal. University of Maryland. College Park, MD 20742

Availability of Coding Based Replication Schemes. Gagan Agrawal. University of Maryland. College Park, MD 20742 Availability of Coding Based Replication Schemes Gagan Agrawal Department of Computer Science University of Maryland College Park, MD 20742 Abstract Data is often replicated in distributed systems to improve

More information

Using Timestamps to Track Causal Dependencies

Using Timestamps to Track Causal Dependencies Using Timestamps to Track Causal Dependencies J. A. David McWha Dept. of Computer Science, University of Waikato, Private Bag 315, Hamilton jadm@cs.waikato.ac.nz ABSTRACT As computer architectures speculate

More information

Process- Concept &Process Scheduling OPERATING SYSTEMS

Process- Concept &Process Scheduling OPERATING SYSTEMS OPERATING SYSTEMS Prescribed Text Book Operating System Principles, Seventh Edition By Abraham Silberschatz, Peter Baer Galvin and Greg Gagne PROCESS MANAGEMENT Current day computer systems allow multiple

More information

SYNCSIM: A SYNCHRONOUS SIMPLE OPTIMISTIC SIMULATION TECHNIQUE BASED ON A GLOBAL PARALLEL HEAP EVENT QUEUE. Sushil K. Prasad Zhiyong Cao

SYNCSIM: A SYNCHRONOUS SIMPLE OPTIMISTIC SIMULATION TECHNIQUE BASED ON A GLOBAL PARALLEL HEAP EVENT QUEUE. Sushil K. Prasad Zhiyong Cao Proceedings of the Winter Simulation Conference S. Chick, P. J. Sánchez, D. Ferrin, and D. J. Morrice, eds. SYNCSIM: A SYNCHRONOUS SIMPLE OPTIMISTIC SIMULATION TECHNIQUE BASED ON A GLOBAL PARALLEL HEAP

More information

An Eæcient Conditional-knowledge based. Optimistic Simulation Scheme. Atul Prakash. June 29, 1991.

An Eæcient Conditional-knowledge based. Optimistic Simulation Scheme. Atul Prakash.   June 29, 1991. An Eæcient Conditional-knowledge based Optimistic Simulation Scheme Atul Prakash Rajalakshmi Subramanian Department of Electrical Engineering and Computer Science University of Michigan, Ann Arbor, MI

More information

Curriculum Vitae of Paolo Romano

Curriculum Vitae of Paolo Romano Curriculum Vitae of Paolo Romano Personal Information Place and Date of Birth: Rome (Italy), 4 March 1979 Citizenship: Italian Office Address: Dipartimento di Informatica e Sistemistica Antonio Ruberti

More information

A taxonomy of race. D. P. Helmbold, C. E. McDowell. September 28, University of California, Santa Cruz. Santa Cruz, CA

A taxonomy of race. D. P. Helmbold, C. E. McDowell. September 28, University of California, Santa Cruz. Santa Cruz, CA A taxonomy of race conditions. D. P. Helmbold, C. E. McDowell UCSC-CRL-94-34 September 28, 1994 Board of Studies in Computer and Information Sciences University of California, Santa Cruz Santa Cruz, CA

More information

AN OBJECT-ORIENTED VISUAL SIMULATION ENVIRONMENT FOR QUEUING NETWORKS

AN OBJECT-ORIENTED VISUAL SIMULATION ENVIRONMENT FOR QUEUING NETWORKS AN OBJECT-ORIENTED VISUAL SIMULATION ENVIRONMENT FOR QUEUING NETWORKS Hussam Soliman Saleh Al-Harbi Abdulkader Al-Fantookh Abdulaziz Al-Mazyad College of Computer and Information Sciences, King Saud University,

More information

Transparently Mixing Undo Logs and Software Reversibility for State Recovery in Optimistic PDES

Transparently Mixing Undo Logs and Software Reversibility for State Recovery in Optimistic PDES Transparently Mixing Undo Logs and Software Reversibility for State Recovery in Optimistic PDES Davide Cingolani, Alessandro Pellegrini, Francesco Quaglia DIAG Sapienza, University of Rome Via Ariosto

More information

Timestamp Representations for Virtual Sequences

Timestamp Representations for Virtual Sequences Timestamp Representations for Virtual equences John G. Cleary, J. A. David McWha, Murray Pearson Dept of Computer cience, University of Waikato, Private Bag 305, Hamilton, New Zealand. {jcleary, jadm,

More information

On Modeling Data Dissemination for LCCIs

On Modeling Data Dissemination for LCCIs On Modeling Data Dissemination for LCCIs Catello Di Martino and Christian Esposito Dipartimento di Informatica e Sistemistica (DIS) Universitá degli studi di Napoli Federico II via Claudio 21, 80125 -

More information

Storage System. Distributor. Network. Drive. Drive. Storage System. Controller. Controller. Disk. Disk

Storage System. Distributor. Network. Drive. Drive. Storage System. Controller. Controller. Disk. Disk HRaid: a Flexible Storage-system Simulator Toni Cortes Jesus Labarta Universitat Politecnica de Catalunya - Barcelona ftoni, jesusg@ac.upc.es - http://www.ac.upc.es/hpc Abstract Clusters of workstations

More information

Database Management Systems Reliability Management

Database Management Systems Reliability Management Database Management Systems Reliability Management D B M G 1 DBMS Architecture SQL INSTRUCTION OPTIMIZER MANAGEMENT OF ACCESS METHODS CONCURRENCY CONTROL BUFFER MANAGER RELIABILITY MANAGEMENT Index Files

More information

TECHNICAL RESEARCH REPORT

TECHNICAL RESEARCH REPORT TECHNICAL RESEARCH REPORT A Resource Reservation Scheme for Synchronized Distributed Multimedia Sessions by W. Zhao, S.K. Tripathi T.R. 97-14 ISR INSTITUTE FOR SYSTEMS RESEARCH Sponsored by the National

More information

CHAPTER 3 RECOVERY & CONCURRENCY ADVANCED DATABASE SYSTEMS. Assist. Prof. Dr. Volkan TUNALI

CHAPTER 3 RECOVERY & CONCURRENCY ADVANCED DATABASE SYSTEMS. Assist. Prof. Dr. Volkan TUNALI CHAPTER 3 RECOVERY & CONCURRENCY ADVANCED DATABASE SYSTEMS Assist. Prof. Dr. Volkan TUNALI PART 1 2 RECOVERY Topics 3 Introduction Transactions Transaction Log System Recovery Media Recovery Introduction

More information

AN ALGORITHM FOR FULLY-REVERSIBLE OPTIMISTIC PARALLEL SIMULATION. Michael D. Peters Christopher D. Carothers

AN ALGORITHM FOR FULLY-REVERSIBLE OPTIMISTIC PARALLEL SIMULATION. Michael D. Peters Christopher D. Carothers Proceedings of the 2003 Winter Simulation Conference S. Chick, P. J. Sánchez, D. Ferrin, and D. J. Morrice, eds. AN ALGORITHM FOR FULLY-REVERSIBLE OPTIMISTIC PARALLEL SIMULATION Michael D. Peters Christopher

More information

System Malfunctions. Implementing Atomicity and Durability. Failures: Crash. Failures: Abort. Log. Failures: Media

System Malfunctions. Implementing Atomicity and Durability. Failures: Crash. Failures: Abort. Log. Failures: Media System Malfunctions Implementing Atomicity and Durability Chapter 22 Transaction processing systems have to maintain correctness in spite of malfunctions Crash Abort Media Failure 1 2 Failures: Crash Processor

More information

Incremental Checkpointing with Application to Distributed Discrete Event Simulation

Incremental Checkpointing with Application to Distributed Discrete Event Simulation Incremental Checkpointing with Application to Distributed Discrete Event Simulation Huining Thomas Feng Edward A. Lee Electrical Engineering and Computer Sciences University of California at Berkeley Technical

More information

Optimistic Message Logging for Independent Checkpointing. in Message-Passing Systems. Yi-Min Wang and W. Kent Fuchs. Coordinated Science Laboratory

Optimistic Message Logging for Independent Checkpointing. in Message-Passing Systems. Yi-Min Wang and W. Kent Fuchs. Coordinated Science Laboratory Optimistic Message Logging for Independent Checkpointing in Message-Passing Systems Yi-Min Wang and W. Kent Fuchs Coordinated Science Laboratory University of Illinois at Urbana-Champaign Abstract Message-passing

More information

On Checkpoint Latency. Nitin H. Vaidya. Texas A&M University. Phone: (409) Technical Report

On Checkpoint Latency. Nitin H. Vaidya. Texas A&M University.   Phone: (409) Technical Report On Checkpoint Latency Nitin H. Vaidya Department of Computer Science Texas A&M University College Station, TX 77843-3112 E-mail: vaidya@cs.tamu.edu Phone: (409) 845-0512 FAX: (409) 847-8578 Technical Report

More information

TIME WARP ON A TRANSPUTER PLATFORM: PILOT STUDY WITH ASYNCHRONOUS CELLULAR AUTOMATA

TIME WARP ON A TRANSPUTER PLATFORM: PILOT STUDY WITH ASYNCHRONOUS CELLULAR AUTOMATA TIME WARP ON A TRANSPUTER PLATFORM: PILOT STUDY WITH ASYNCHRONOUS CELLULAR AUTOMATA B. J. Overeinder, P. M. A. Sloot, and L. O. Hertzberger High Performance Computing Group. Department of Computer Systems,

More information

Egemen Tanin, Tahsin M. Kurc, Cevdet Aykanat, Bulent Ozguc. Abstract. Direct Volume Rendering (DVR) is a powerful technique for

Egemen Tanin, Tahsin M. Kurc, Cevdet Aykanat, Bulent Ozguc. Abstract. Direct Volume Rendering (DVR) is a powerful technique for Comparison of Two Image-Space Subdivision Algorithms for Direct Volume Rendering on Distributed-Memory Multicomputers Egemen Tanin, Tahsin M. Kurc, Cevdet Aykanat, Bulent Ozguc Dept. of Computer Eng. and

More information

A Wait-free Multi-word Atomic (1,N) Register for Large-scale Data Sharing on Multi-core Machines

A Wait-free Multi-word Atomic (1,N) Register for Large-scale Data Sharing on Multi-core Machines A Wait-free Multi-word Atomic (1,N) Register for Large-scale Data Sharing on Multi-core Machines Mauro Ianni, Alessandro Pellegrini DIAG Sapienza Università di Roma, Italy Email: {mianni,pellegrini}@dis.uniroma1.it

More information

An Efficient Execution Scheme for Designated Event-based Stream Processing

An Efficient Execution Scheme for Designated Event-based Stream Processing DEIM Forum 2014 D3-2 An Efficient Execution Scheme for Designated Event-based Stream Processing Yan Wang and Hiroyuki Kitagawa Graduate School of Systems and Information Engineering, University of Tsukuba

More information

This paper describes and evaluates the Dual Reinforcement Q-Routing algorithm (DRQ-Routing)

This paper describes and evaluates the Dual Reinforcement Q-Routing algorithm (DRQ-Routing) DUAL REINFORCEMENT Q-ROUTING: AN ON-LINE ADAPTIVE ROUTING ALGORITHM 1 Shailesh Kumar Risto Miikkulainen The University oftexas at Austin The University oftexas at Austin Dept. of Elec. and Comp. Engg.

More information

Numerical approach estimate

Numerical approach estimate Simulation Nature of simulation Numericalapproachfor investigating models of systems. Data are gathered to estimatethe true characteristics of the model. Garbage in garbage out! One of the techniques of

More information

Transaction Management. Pearson Education Limited 1995, 2005

Transaction Management. Pearson Education Limited 1995, 2005 Chapter 20 Transaction Management 1 Chapter 20 - Objectives Function and importance of transactions. Properties of transactions. Concurrency Control Deadlock and how it can be resolved. Granularity of

More information

Implementing Sequential Consistency In Cache-Based Systems

Implementing Sequential Consistency In Cache-Based Systems To appear in the Proceedings of the 1990 International Conference on Parallel Processing Implementing Sequential Consistency In Cache-Based Systems Sarita V. Adve Mark D. Hill Computer Sciences Department

More information