More for your money
Published 14 October 2012
Benjamin Farley, Ari Juels, Venkatanathan Varadarajan, Thomas Ristenpart, Kevin D. Bowers, Michael M. Swift
Citations168
Generate an AI Snapshot to get a quick, structured summary of this paper.
Study Snapshot
ObjectiveStudy objective
MethodsResearch methodology
PopulationPopulation studied
Sample sizeSample sizes
OutcomesStudy outcomes here
ResultsStudy results comes here
LimitationsResearch study limitations comes here
A concise AI-generated summary of the paper will appear here once you click Generate AI Snapshot.
TL;DR
This work confirms the (oft-reported) performance differences between supposedly identical instances, and leads to identify fruitful targets for placement gaming, such as CPU, network, and storage performance, and develops a formal model for placement strategies and evaluates potential strategies via simulation.
Abstract
Infrastructure-as-a-system compute clouds such as Amazon's EC2 allow users to pay a flat hourly rate to run their virtual machine (VM) on a server providing some combination of CPU access, storage, and network. But not all VM instances are created equal: distinct underlying hardware differences, contention, and other phenomena can result in vastly differing performance across supposedly equivalent instances. The result is striking variability in the resources received for the same price.
Keywords
Computer Science
MIT Press eBooksIntroduction to Reinforcement Learning
6,868 Citations1998Richard S. Sutton, Andrew G. Barto
In Reinforcement Learning, Richard Sutton and Andrew Barto provide a clear and simple account of the key ideas and algorithms of reinforcement learning.
Hey, you, get off of my cloud
2,015 Citations2009Thomas Ristenpart, Eran Tromer +2 more
ACM SIGARCH Computer Architecture NewsSPEC CPU2006 benchmark descriptions
1,973 Citations2006John L. Henning
On August 24, 2006, the Standard Performance Evaluation Corporation (SPEC) announced CPU2006, which replaces CPU2000, and the SPEC CPU benchmarks are widely used in both industry and academia.
UC BerkeleyImproving MapReduce performance in heterogeneous environments
1,619 Citations2008Matei Zaharia, Andy Konwinski +3 more
A new scheduling algorithm, Longest Approximate Time to End (LATE), that is highly robust to heterogeneity and can improve Hadoop response times by a factor of 2 in clusters of 200 virtual machines on EC2.
CloudCmp
874 Citations2010Ang Li, Xiaowei Yang +2 more
Applying CloudCmp to four cloud providers that together account for most of the cloud customers today, it is found that their offered services vary widely in performance and costs, underscoring the need for thoughtful provider selection.
Proceedings of the VLDB EndowmentRuntime measurements in the cloud
674 Citations2010Jörg Schad, Jens Dittrich +1 more
A study of the performance variance of the most widely used Cloud infrastructure (Amazon EC2) from different perspectives using established microbenchmarks to measure performance variance in CPU, I/O, and network and a multi-node MapReduce application to quantify the impact on real dataintensive applications.
The Impact of Virtualization on Network Performance of Amazon EC2 Data Center
660 Citations2010Guohui Wang, T. S. Eugene Ng
The results show that even though the data center network is lightly utilized, virtualization can still cause significant throughput instability and abnormal delay variations.
Biomedical named entity recognition using conditional random fields and rich feature sets
527 Citations2004Burr Settles
As the wealth of biomedical knowledge in the form of literature increases, there is a rising need for effective natural language processing tools to assist in organizing, curating, and retrieving this information.
On the Performance Variability of Production Cloud Services
328 Citations2011Alexandru Iosup, Nezih Yigitbasi +1 more
It is found that the performance of about half of the cloud services investigated exhibits yearly and daily patterns, but also that most services have periods of especially stable performance, which gives evidence that performance variability can be an important factor in cloud provider selection.
Learning from Bullying Traces in Social Media
306 Citations2012Junming Xu, Kwang-Sung Jun +2 more
Evidence is presented that social media, with appropriate natural language processing techniques, can be a valuable and abundant data source for the study of bullying in both worlds.
Named entity recognition with character-level models
195 Citations2003Dan Klein, Joseph Smarr +2 more
Two named-entity recognition models which use characters and character n-grams either exclusively or as an important part of their data representation are discussed, both of which are a character-level HMM with minimal context information and a maximum-entropy conditional markov model with substantially richer context features.
Empirical evaluation of latency-sensitive application performance in the cloud
183 Citations2010Sean Barker, Prashant Shenoy
The results reveal that the jitter and the throughput seen by a latency-sensitive application can indeed degrade due to background load from other virtual machines and that careful configuration of the resource control mechanisms within the virtualization layer can mitigate, but not eliminate, this interference.
Lecture notes in computer scienceEC2 Performance Analysis for Resource Provisioning of Service-Oriented Applications
168 Citations2010Jiang Dejun, Guillaume Pierre +1 more
It is demonstrated that the performance of virtual instances in Amazon EC2 is relatively stable over time with fluctuations of mean response time within at most 8% of the longterm average.
A comprehensive scheduler for asymmetric multicore systems
140 Citations2010Juan Carlos Sáez, Manuel Prieto +2 more
CAMP, a Comprehensive AMP scheduler is proposed, implemented, and evaluated, which delivers both efficiency and TLP specialization, and a new light-weight technique for discovering which threads utilize fast cores most efficiently is proposed.
Seawall: performance isolation for cloud datacenter networks
138 Citations2010Alan Shieh, Srikanth Kandula +2 more
Seawall is proposed, an edge-based solution, that achieves max-min fairness across tenant VMs by sending traffic through congestion-controlled, hypervisor-to-hypervisor tunnels.
Exploiting hardware heterogeneity within the same instance type of Amazon EC2
93 Citations2012Zhonghong Ou, Hao Zhuang +3 more
This paper looks into the hardware heterogeneity and the corresponding performance variation within the same instance type of Amazon Elastic Compute Cloud (Amazon EC2), finding out that there exist three different subtypes of hardware configuration in the standard large instance.
IEEE Transactions on Automatic ControlMulti-armed bandits with switching penalties
84 Citations1996M. Asawa, Demosthenis Teneketzis
It is shown that the aforementioned property of optimal policies is sufficient to determine an optimal allocation strategy, and sufficient conditions for optimality of allocation strategies, based on limited look-ahead techniques, are established.
CRUISE
76 Citations2012Aamer Jaleel, Hashem H. Najaf-abadi +3 more
This study presents a detailed analysis on the interactions between intelligent scheduling and smart cache replacement policies and proposes Cache Replacement and Utility-aware Scheduling (CRUISE)-a hardware/software co-designed approach for shared cache management.
ComputerUsing Mathematical Modeling in Provisioning a Heterogeneous Cloud Computing Environment
74 Citations2011Sungkap Yeo, Hsien-Hsin S. Lee
To achieve the optimal performance per utility, the slowest node's response time should be no more than three times that of the fastest node.
Exploring the Performance Fluctuations of HPC Workloads on Clouds
63 Citations2010Yaakoub El-Khamra, Hyunjoo Kim +2 more
This work finds strong fluctuations from the mean performance on EC2 and Eucalyptus-based cloud systems and concludes that variations in communication times account for the bulk of the experiment-to-experiment fluctuations of the performance.
USENIX conference on Web application developmentResource provisioning of web applications in heterogeneous clouds
60 Citations2011Jiang Dejun, Guillaume Pierre +1 more
Initial Findings for Provisioning Variation in Cloud Computing
56 Citations2010M. Suhail Rehman, Majd Sakr
The initial findings indicate that provisioning variation can impact performance by a factor of 5 primarily due to I/O contention.
Big Data versus the Crowd: Looking for Relationships in All the Right Places
46 Citations2012Ce Zhang, Feng Niu +2 more
The experiments show that increasing the corpus size for distant supervision has a statistically significant, positive impact on quality (F1 score), and human feedback has a positive and statistically significant; but lower, impact on precision and recall.
USENIX conference on Hot Topics in Cloud CcomputingThe seven deadly sins of cloud computing research
45 Citations2012Malte Schwarzkopf, Derek G. Murray +1 more
This paper identifies and discusses seven "deadly sins" of distributed parallelism, present evidence illustrating that they pose real problems, and discuss ways for the community to avoid them in the future.
IEEE International Conference on Cloud Computing Technology and ScienceMigration, assignment, and scheduling of jobs in virtualized environment
29 Citations2011Seung-Hwan Lim, Jaeseok Huh +2 more
A method to analyze the migration time and the performance impact on multiresource shared systems for completing given VM assignment plan is proposed and may contribute to create more robust performance in virtualized environment.