Filters








90 Hits in 6.2 sec

Learning to Dispatch Multi-Server Jobs in Bipartite Graphs with Unknown Service Rates [article]

Hailiang Zhao, Shuiguang Deng, Feiyi Chen, Jianwei Yin, Schahram Dustdar, Albert Y. Zomaya
2022 arXiv   pre-print
In this paper, we propose a dispatching algorithm for multi-server jobs that learns the unknown service rates and simultaneously maximizes the expected Accumulative Social Welfare (Asw).  ...  However, current job dispatching algorithms require the service rates to be changeless and knowable, which is difficult to realize in production systems.  ...  • We formally formulate the dispatching problem for multi-server jobs in a bipartite graph with unknown service rates. The objective is to maximize the EASW from a long-term vision.  ... 
arXiv:2204.04371v1 fatcat:qm5u356ryfbmvjqx2cofpgvcvi

Distributed Dispatching in the Parallel Server Model

Guy Goren, Shay Vargaftik, Yoram Moses, Hagit Attiya
2020 International Symposium on Distributed Computing  
With the rapid increase in the size and volume of cloud services and data centers, architectures with multiple job dispatchers are quickly becoming the norm.  ...  Nevertheless, current solutions to load balancing in such systems admit a paradoxical behavior in which more accurate information regarding server queue lengths degrades performance due to herding and  ...  (The distributions of arrival rates at dispatchers are unknown but assumed to be the same, as are those of server processing rates.)  ... 
doi:10.4230/lipics.disc.2020.14 dblp:conf/wdag/GorenVM20 fatcat:ehf7wq6upfdcja565rqr753aia

Distributed Dispatching in the Parallel Server Model [article]

Guy Goren, Shay Vargaftik, Yoram Moses
2020 arXiv   pre-print
With the rapid increase in the size and volume of cloud services and data centers, architectures with multiple job dispatchers are quickly becoming the norm.  ...  Nevertheless, current solutions to load balancing in such systems admit a paradoxical behavior in which more accurate information regarding server queue lengths degrades performance due to herding and  ...  (The distributions of arrival rates at dispatchers are unknown but assumed to be the same, as are those of server processing rates.)  ... 
arXiv:2008.00793v1 fatcat:iivtpwjcube25l2cg3y2atqcgu

Theoretically Guaranteed Online Workload Dispatching for Deadline-Aware Multi-Server Jobs [article]

Hailiang Zhao, Shuiguang Deng, Jianwei Yin, Schahram Dustdar, Albert Y. Zomaya
2022 arXiv   pre-print
Multi-server jobs are imperative in modern computing clusters.  ...  Efficient online workload dispatching is crucial but challenging to co-located heterogeneous multi-server jobs.  ...  To fill the theoretical gap, in this paper, we study a general online workload dispatching problem for deadline-aware multi-server jobs.  ... 
arXiv:2112.02456v3 fatcat:sskh5kwybjeutbwit5hyxbqfym

Learning-NUM: Network Utility Maximization with Unknown Utility Functions and Queueing Delay [article]

Xinzhe Fu, Eytan Modiano
2020 arXiv   pre-print
In this paper, we propose a new NUM framework, Learning-NUM, where the users' utility functions are unknown apriori and the utility function values of the traffic rates can be observed only after the corresponding  ...  Finally, to demonstrate the practical applicability of the Learning-NUM framework, we apply it to three application scenarios including database query, job scheduling and video streaming.  ...  Job Scheduling Consider a discrete-time system with with a set of job schedulers (dispatchers) { 1 , . . . , } and a set of parallel servers { 1 , . . . , } that form a bipartite graph.  ... 
arXiv:2012.09222v1 fatcat:76h3lq6ixjeprjqdolpcwcxz7u

An SMDP-Based Approach to Thermal-Aware Task Scheduling in NoC-based MPSoC platforms [article]

Farnaz Niknia, Kiamehr Rezaee, Vesal Hakami
2020 arXiv   pre-print
One efficient approach to control chip-wide thermal distribution in multi-core systems is the optimization of online assignments of tasks to processing cores.  ...  Compared to related research, the simulation results show a nearly 6 Kelvin reduction in system average peak temperature and 66 milliseconds decrease in mean task service time.  ...  Given the specifications of our system model, the arrival queue corresponds to an infinite-length multi-server queue with a Poisson arrival process, and exponentially distributed service time.  ... 
arXiv:2009.02813v1 fatcat:puuf6nkvzrhsdpj65qz2f6xa2u

AntMan: Dynamic Scaling on GPU Clusters for Deep Learning

Wencong Xiao, Shiru Ren, Yong Li, Yang Zhang, Pengyang Hou, Zhi Li, Yihui Feng, Wei Lin, Yangqing Jia
2020 USENIX Symposium on Operating Systems Design and Implementation  
This paper presents AntMan, a deep learning infrastructure that co-designs cluster schedulers with deep learning frameworks and has been deployed in production at Alibaba to manage tens of thousands of  ...  Evaluations show that AntMan improves the overall GPU memory utilization by 42% and computation utilization by 34% in our multi-tenant cluster without compromising fairness, presenting a new approach to  ...  We would also like to thank Chen Xing, Jin Ouyang, Xinyuan Li, Lixue Xia for their help in improving quality of writing.  ... 
dblp:conf/osdi/XiaoRLZHLFLJ20 fatcat:lfa2xrj7zveulikmjxzi66e7fm

Reinforcement Learning-Empowered Mobile Edge Computing for 6G Edge Intelligence [article]

Peng Wei, Kun Guo, Ye Li, Jue Wang, Wei Feng, Shi Jin, Ning Ge, Ying-Chang Liang
2022 arXiv   pre-print
Thanks to the evolved reinforcement learning (RL), upon iteratively interacting with the dynamic and random environment, its trained agent can intelligently obtain the optimal policy in MEC.  ...  Finally, the open challenges are discussed to provide helpful guidance for future research in RL training and learning MEC.  ...  Attributed to the mixed-integer problem, the DDPG algorithm is employed with the mapping from continuous-valued actions to refined discrete-valued actions by the constructed bipartite graph.  ... 
arXiv:2201.11410v4 fatcat:24igkq4kbrb2pjzwf3mf3n7qtq

Big Data Analytics [article]

Ahmed Masmoudi
2017 Zenodo  
The content of the book is to be used by the students for the reference for the subject "Big Data Analytics"  ...    Finding Complete Bipartite Subgraphs Taking a large bipartite graph G , and to find instances of Ks,t within it.  ...  We denote this graph by Ks,t.  Draw an analogy between complete bipartite graphs as subgraphs of general bipartite graphs and cliques as subgraphs of general graphs.  ... 
doi:10.5281/zenodo.573349 fatcat:qg7licyavbgbtph6jadfm6bncu

VNF and Container Placement: Recent Advances and Future Trends [article]

Wissal Attaoui, Essaid Sabir, Halima Elbiaze, Mohsen Guizani
2022 arXiv   pre-print
Virtualization is not limited to simply replacing physical machines with virtual machines or VNFs, but may also include micro-services, containers, and cloud-native systems.  ...  This decoupling allows network services, referred to as Virtualized Network Functions (VNFs), to be hosted on commodity hardware which simplifies and enhances service deployment and management for providers  ...  The main objective relies on maximum matching with minimum bipartite graph (BG) cost.  ... 
arXiv:2204.00178v1 fatcat:giwlibsaknbkdnuvkao6nhkdnq

Reinforcement Learning-Empowered Mobile Edge Computing for 6G Edge Intelligence

Peng Wei, Kun Guo, Ye Li, Jue Wang, Wei Feng, Shi Jin, Ning Ge, Ying-Chang Liang
2022 IEEE Access  
Thanks to the evolved reinforcement learning (RL), upon iteratively interacting with the dynamic and random environment, its trained agent can intelligently obtain the optimal policy in MEC.  ...  Finally, the open challenges are discussed to provide helpful guidance for future research in RL training and learning MEC.  ...  Attributed to the mixed-integer problem, the DDPG algorithm is employed with the mapping from continuous-valued actions to refined discrete-valued actions by the constructed bipartite graph.  ... 
doi:10.1109/access.2022.3183647 fatcat:pd5z6q4innd5jl25g4r7b4nq3i

Ridesourcing systems: A framework and review

Hai Wang, Hai Yang
2019 Transportation Research Part B: Methodological  
service), etc., to more than 550 million users in 400 cities in China, with 30 million daily trips as of mid-2019 (see DMR, 2019b ).  ...  After each trip, passengers can rate the drivers who provided the transportation service, which helps to quantify the quality of service provided by the affiliated drivers.  ...  We also express our sincere appreciation to the seven anonymous referees for their invaluable comments and suggestions.  ... 
doi:10.1016/j.trb.2019.07.009 fatcat:m5rxzhsxsfaunoep77nsqdhbny

Auto-scaling Web Applications in Clouds: A Taxonomy and Survey [article]

Chenhao Qu, Rodrigo N. Calheiros, Rajkumar Buyya
2017 arXiv   pre-print
under a dynamic workload to minimize resource cost while satisfying Quality of Service (QoS) requirements.  ...  We present a taxonomy of auto-scalers according to the identified challenges and key properties. We analyze the surveyed works and map them to the taxonomy to identify the weaknesses in this field.  ...  Yaser Mansouri, Xunyun Liu, Minxian Xu, and Bowen Zhou for their valuable comments and suggestions in improving the quality of the paper.  ... 
arXiv:1609.09224v6 fatcat:dkk2ftpvpbcnvhcmc6lz2omwa4

A survey on data center networking for cloud computing

Bin Wang, Zhengwei Qi, Ruhui Ma, Haibing Guan, Athanasios V. Vasilakos
2015 Computer Networks  
In recent years, the growing importance of data center networking has drawn much attention to related issues including connective simplification and service stability.  ...  In our attempt to build insight relevant to future research, we also present some open research issues.  ...  A central controller determines the desired bandwidth and path for each VM by building a bipartite graph between the VDC and the physical infrastructure.  ... 
doi:10.1016/j.comnet.2015.08.040 fatcat:om742vsrd5bjrb5nyaqrqlxrze

Dynamic Assignment Control of a Closed Queueing Network under Complete Resource Pooling [article]

Siddhartha Banerjee, Yash Kanoria, Pengyu Qian
2020 arXiv   pre-print
(ii) Service provider selection in scrip systems (like for babysitting or for kidney exchange): With only cosmetic modifications to the setup, our results translate fully to a model of scrip systems and  ...  We study the design of dynamic assignment control in networks with a fixed number of circulating resources (supply units).  ...  The waste of service token can be interpreted as the server starting to serve a "dummy job". Service of dummy jobs corresponds to server idleness in the "classical" model.  ... 
arXiv:1803.04959v3 fatcat:uj7lo77l35cb5b5ktsqwc2ldva
« Previous Showing results 1 — 15 out of 90 results