Filters








41,959 Hits in 6.0 sec

On Choosing a Task Assignment Policy for a Distributed Server System

Mor Harchol-Balter, Mark E. Crovella, Cristina D. Murta
1999 Journal of Parallel and Distributed Computing  
We consider a distributed server system model and ask which policy should be used for assigning tasks to hosts.  ...  We consider four task assignment policies commonly proposed for such distributed server systems: Round-Robin, Random, Size-Based, in which all tasks within a give size range are assigned to a particular  ...  Given the above model of a distributed server system, we ask how to select the best task assignment policy.  ... 
doi:10.1006/jpdc.1999.1577 fatcat:gjot3tnqgjgmxnuoyarya6rlyy

On Choosing a Task Assignment Policy for a Distributed Server System [chapter]

Mor Harchol-Balter, Mark E. Crovella, Cristina D. Murta
1998 Lecture Notes in Computer Science  
We consider a distributed server system model and ask which policy should be used for assigning tasks to hosts.  ...  We consider four task assignment policies commonly proposed for such distributed server systems: Round-Robin, Random, Size-Based, in which all tasks within a give size range are assigned to a particular  ...  Given the above model of a distributed server system, we ask how to select the best task assignment policy.  ... 
doi:10.1007/3-540-68061-6_19 fatcat:eyf4k3zhgvbd5jxhknyr76iea4

To split or not to split: Selecting the right server with batch arrivals

Esa Hyytiä, Samuli Aalto
2013 Operations Research Letters  
We consider a dispatching system, where jobs, arriving in batches, are assigned to single-server FCFS queues. Batches can be split to different queues on per job basis.  ...  By using the first policy improvement step of the MDP framework, we are able to derive robust dispatching policies, which split arriving batches only when deemed advantageous.  ...  In this paper, we have addressed the joint-problem of splitting the tasks and choosing the server for each sub-task, i.e., the task assignment problem with batch arrivals.  ... 
doi:10.1016/j.orl.2013.03.007 fatcat:7wb56r22lnapxdg6i2eronfgyu

The design of a real-time event manager component

D Masson, S Midonnet
2010 2010 10th Annual International Conference on New Technologies of Distributed Systems (NOTERE)  
This component manages the relations between the triggering of distributed events, the events handlers and the resource allocations (servers and processors).  ...  The feasibility analysis is processed at a component level and takes into account the events relations and the association of the servers with the processors (affinity).  ...  And when several handlers are assigned to the same task server, their interference on the other tasks in the system are linked.  ... 
doi:10.1109/notere.2010.5536621 dblp:conf/notere/MassonM10 fatcat:kvp7sdyi5jfhvijjsu6to4ux3q

On the relative value of local scheduling versus routing in parallel server systems

Rong Wu, D.G. Down
2007 2007 International Conference on Parallel and Distributed Systems  
We consider a system with a dispatcher and several identical servers in parallel. Task processing times are known upon arrival. We first study the impact of the local scheduling policy at a server.  ...  Our numerical results show that the performance (mean waiting time) of such a policy could be significantly better than the best known suggested policies that use FCFS at each server.  ...  For a single server system, Schrage in [12] showed that for an arbitrary arrival process and arbitrary processing time distribution, SRPT minimizes the number of tasks in system.  ... 
doi:10.1109/icpads.2007.4447751 dblp:conf/icpads/WuD07 fatcat:6fw6tfjtireibmltcd5m4c6m4e

A Comprehensive Study of Multi-Scheduling Schemes Performance

Xiao Chen, Lei Jiang, Jin Wang
2015 International Journal of u- and e- Service, Science and Technology  
The research was started regarding the simulation of a novel scheduling policy (Task Assignment by Guessing Size) associated with other two previous task assignment policies (Random and JSQ).  ...  Today distributed server systems have been widely used in many areas because they enhance the computing power while being cost-effective and more efficient.  ...  Acknowledgements This paper is a revised and expanded version of a paper entitled "Performance Evaluation of Multi-Scheduling Schemes" presented at CST 2015, Suzhou, China, April  ... 
doi:10.14257/ijunesst.2015.8.9.06 fatcat:6ujwaa75bffolkezcdewz5py3q

Evaluation of Task Assignment Policies for Supercomputing Servers: The Case for Load Unbalancing and Fairness

Bianca Schroeder, Mor Harchol-Balter
2004 Cluster Computing  
The biggest question which arises in such distributed server systems is what is a good rule for assigning jobs to host machines: i.e. what is a good task assignment policy.  ...  This alternative setup may be described simply as a collection of multiprocessors or a distributed server system.  ...  Many thanks to the system administrators who spoke with us about task assignment in their systems, in particular Scott Blomquist and Chris Hill.  ... 
doi:10.1023/b:clus.0000018564.05723.a2 fatcat:xh3fpx4jmnh3hgxnkjzm5qlbby

Task assignment with unknown duration

Mor Harchol-Balter
2002 Journal of the ACM  
We consider a distributed server system and ask which policy should be used for assigning tasks to hosts. In our server, tasks are not preemptible.  ...  We analyze several natural task assignment policies and propose a new one TAGS Task Assignment based on Guessing Size.  ...  Designing a distributed server system often comes down to choosing the best" task assignment policy for the given model and user requirements.  ... 
doi:10.1145/506147.506154 fatcat:c7dvtyai6nfqxhz5qcvejaxvna

On Delay-Optimal Scheduling in Queueing Systems with Replications [article]

Yin Sun, C. Emre Koksal, Ness B. Shroff
2017 arXiv   pre-print
One common solution for straggler mitigation is to replicate a task on multiple servers and wait for one replica of the task to finish early.  ...  In this paper, we present a comprehensive study on delay-optimal scheduling of replications in both centralized and distributed multi-server systems.  ...  Task Replication: When a task is assigned, it is replicated on all m servers. 2.  ... 
arXiv:1603.07322v8 fatcat:iacxmhmozbcu3cgqqhzicjhs2q

Rosella: A Self-Driving Distributed Scheduler for Heterogeneous Clusters [article]

Qiong Wu, Zhenming Liu
2021 arXiv   pre-print
We present Rosella, a new self-driving, distributed approach for task scheduling in heterogeneous clusters.  ...  Large-scale interactive web services and advanced AI applications make sophisticated decisions in real-time, based on executing a massive amount of computation tasks on thousands of servers.  ...  The SQ(2) policy would chooses the one server with the shorter queue while the LL(2) policy would chooses the server with shorter waiting time.  ... 
arXiv:2010.15206v3 fatcat:a6h4ddkelnahlfkujgqiaanhzu

Scheduling and Priority Mapping for Static Real-Time Middleware [chapter]

Lisa Cingiser Dipippo, Victor Fay Wolfe, Levon Esibov, Gregory Cooper, Ramachandra Bethmangalkar, Russell Johnston, Bhavani Thuraisingham, John Mauer
2001 Challenges in Design and Implementation of Middlewares for Real-Time Systems  
The technique uses global deadline monotonic priority assignment to clients and the Distributed Priority Ceiling protocol to provide concurrency control and priorities for server execution.  ...  This paper presents a middleware real-time scheduling technique for static, distributed, real-time applications.  ...  priority inheritance for requests that are queued on a server. 47 Distributed Scheduling Scheduling tasks in a distributed system involves assigning priorities to the tasks and controlling access  ... 
doi:10.1007/978-1-4615-1695-8_3 fatcat:sga7wjrdbbenpc6guszeuptlsy

Offload (only) the right jobs: Robust offloading using the Markov decision processes

Esa Hyytia, Thrasyvoulos Spyropoulos, Jorg Ott
2015 2015 IEEE 16th International Symposium on A World of Wireless, Mobile and Multimedia Networks (WoWMoM)  
Furthermore, for both type (ii) and (iii) tasks, there is typically two ways to access the cloud: via a (costly) cellular connection or via intermittently available WLAN hotspots.  ...  For type (iii) tasks, it is of interest to consider when they should be processed locally and when in the cloud.  ...  This boils down to two kinds of dynamic task assignment decisions as illustrated in Fig. 3 : (i) choosing the cloud and the access link, or local processing for the flexible tasks, and (ii) choosing the  ... 
doi:10.1109/wowmom.2015.7158127 dblp:conf/wowmom/HyytiaSO15 fatcat:24vxflwqlvf7hho2juklmz7p4y

RapidSched: static scheduling and analysis for real-time CORBA

Wolfe, DiPippo, Bethmag, Cooper, Johnston, Kortmann, Watson, Wohlever
1999 Engineering Solutions for the Next Millennium 1999 IEEE Canadian Conference on Electrical and Computer Engineering (Cat No 99TH8411) CCECE-99  
RapidSched uses a global, distributed deadline monotonic priority assignment, enforcement of priorities by commercial real-time operating systems, and distributed priority ceiling resource management.  ...  This paper presents a real-time CORBA Scheduling Service called RapidSched.  ...  RapidSched uses a global deadline monotonic priority assignment technique along with a distributed priority ceiling resource access protocol for fixed priority static distributed systems.  ... 
doi:10.1109/words.1999.806558 dblp:conf/words/WolfeDBCJKWW99 fatcat:pie2poz5jvar5bjppszga5nonq

Round-Robin Routing Policy: Value Functions and Mean Performance with Job- and Server-specific Costs

Esa Hyytia, Samuli Aalto
2014 Proceedings of the 7th International Conference on Performance Evaluation Methodologies and Tools  
The results are demonstrated in the numerical examples, where we compute near optimal task assignment policies for a sample system with two servers.  ...  Keywords Round-Robin, M/G/m-RR, Erl/G/1 queue, Task assignment, Dispatching, Parallel queues, MDP  ...  INTRODUCTION In the task assignment (or routing) problems, one chooses a server for each new job immediately upon the arrival.  ... 
doi:10.4108/icst.valuetools.2013.254379 dblp:conf/valuetools/HyytiaA13 fatcat:4tnzgedlfrh3vjxt7d2dxbjj54

Special Issue on Software Support for Distributed Computing

Ishfaq Ahmad, Francis Lau
1999 Journal of Parallel and Distributed Computing  
Many wide-area distributed systems that are in the making, the Globe system [6] for example, have worldwide scalability as one of their main goals.  ...  The Internet as one gigantic distributed system is no longer a farfetched idea of the visionaries; it is now being used by some to try to solve really tough problems to search for extraterrestrial intelligence  ...  We thank Professor Kai Hwang for his support for the special issue and for continued guidance during the editorial process.  ... 
doi:10.1006/jpdc.1999.1579 fatcat:dzeqwtta4zeeveqvwl3e6hq3um
« Previous Showing results 1 — 15 out of 41,959 results