S. Khan, Donghyuk Lee, Yoongu Kim, Alaa R. Alameldeen, C. Wilkerson, O. Mutlu
As DRAM cells continue to shrink, they become more susceptible to retention failures. DRAM cells that permanently exhibit short retention times are fairly easy to identify and repair through the use of memory tests and row and column redundancy. However, the retention time of many cells may vary over time due to a property called Variable Retention Time (VRT). Since these cells intermittently transition between failing and non-failing states, they are particularly difficult to identify through memory tests alone. In addition, the high temperature packaging process may aggravate this problem as the susceptibility of cells to VRT increases after the assembly of DRAM chips. A promising alternative to manufacture-time testing is to detect and mitigate retention failures after the system has become operational. Such a system would require mechanisms to detect and mitigate retention failures in the field, but would be responsive to retention failures introduced after system assembly and could dramatically reduce the cost of testing, enabling much longer tests than are practical with manufacturer testing equipment. In this paper, we analyze the efficacy of three common error mitigation techniques (memory tests, guardbands, and error correcting codes (ECC)) in real DRAM chips exhibiting both intermittent and permanent retention failures. Our analysis allows us to quantify the efficacy of recent system-level error mitigation mechanisms that build upon these techniques. We revisit prior works in the context of the experimental data we present, showing that our measured results significantly impact these works' conclusions. We find that mitigation techniques that rely on run-time testing alone [38, 27, 50, 26] are unable to ensure reliable operation even after many months of testing. Techniques that incorporate ECC[4, 52], however, can ensure reliable DRAM operation after only a few hours of testing. For example, VS-ECC[4], which couples testing with variable strength codes to allocate the strongest codes to the most error-prone memory regions, can ensure reliable operation for 10 years after only 19 minutes of testing. We conclude that the viability of these mitigation techniques depend on efficient online profiling of DRAM performed without disrupting system operation.
{"title":"The efficacy of error mitigation techniques for DRAM retention failures: a comparative experimental study","authors":"S. Khan, Donghyuk Lee, Yoongu Kim, Alaa R. Alameldeen, C. Wilkerson, O. Mutlu","doi":"10.1145/2591971.2592000","DOIUrl":"https://doi.org/10.1145/2591971.2592000","url":null,"abstract":"As DRAM cells continue to shrink, they become more susceptible to retention failures. DRAM cells that permanently exhibit short retention times are fairly easy to identify and repair through the use of memory tests and row and column redundancy. However, the retention time of many cells may vary over time due to a property called Variable Retention Time (VRT). Since these cells intermittently transition between failing and non-failing states, they are particularly difficult to identify through memory tests alone. In addition, the high temperature packaging process may aggravate this problem as the susceptibility of cells to VRT increases after the assembly of DRAM chips. A promising alternative to manufacture-time testing is to detect and mitigate retention failures after the system has become operational. Such a system would require mechanisms to detect and mitigate retention failures in the field, but would be responsive to retention failures introduced after system assembly and could dramatically reduce the cost of testing, enabling much longer tests than are practical with manufacturer testing equipment.\u0000 In this paper, we analyze the efficacy of three common error mitigation techniques (memory tests, guardbands, and error correcting codes (ECC)) in real DRAM chips exhibiting both intermittent and permanent retention failures. Our analysis allows us to quantify the efficacy of recent system-level error mitigation mechanisms that build upon these techniques. We revisit prior works in the context of the experimental data we present, showing that our measured results significantly impact these works' conclusions. We find that mitigation techniques that rely on run-time testing alone [38, 27, 50, 26] are unable to ensure reliable operation even after many months of testing. Techniques that incorporate ECC[4, 52], however, can ensure reliable DRAM operation after only a few hours of testing. For example, VS-ECC[4], which couples testing with variable strength codes to allocate the strongest codes to the most error-prone memory regions, can ensure reliable operation for 10 years after only 19 minutes of testing. We conclude that the viability of these mitigation techniques depend on efficient online profiling of DRAM performed without disrupting system operation.","PeriodicalId":306456,"journal":{"name":"Measurement and Modeling of Computer Systems","volume":"20 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2014-06-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127515277","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Jinyoung Han, Daejin Choi, Byung-Gon Chun, T. Kwon, Hyunchul Kim, Yanghee Choi
Pinterest, a popular social curating service where people collect, organize, and share content (pins in Pinterest), has gained great attention in recent years. Despite the increasing interest in Pinterest, little research has paid attention to how people collect, manage, and share pins in Pinterest. In this paper, to shed insight on such issues, we study the following questions. How do people collect and manage pins by their tastes in Pinterest? What factors do mainly drive people to share their pins in Pinterest? How do the characteristics of users (e.g., gender, popularity, country) or properties of pins (e.g., category, topic) play roles in propagating pins in Pinterest? To answer these questions, we have conducted a measurement study on patterns of pin curating and sharing in Pinterest. By keeping track of all the newly posted and shared pins in each category (e.g., animal, kids, women's fashion) from June 5 to July 18, 2013, we built 350 K pin propagation trees for 3 M users. With the dataset, we investigate: (1) how users collect and curate pins, (2) how users share their pins and why, and (3) how users are related by shared pins of interest. Our key finding is that pin propagation in Pinterest is mostly driven by pin's properties like its topic, not by user's characteristics like her number of followers. We further show that users in the same community in the interest graph (i.e., representing the relations among users) of Pinterest share pins (i) in the same category with 94% probability and (ii) of the same URL where pins come from with 89% probability. Finally, we explore the implications of our findings for predicting how pins are shared in Pinterest.
{"title":"Collecting, organizing, and sharing pins in pinterest: interest-driven or social-driven?","authors":"Jinyoung Han, Daejin Choi, Byung-Gon Chun, T. Kwon, Hyunchul Kim, Yanghee Choi","doi":"10.1145/2591971.2591996","DOIUrl":"https://doi.org/10.1145/2591971.2591996","url":null,"abstract":"Pinterest, a popular social curating service where people collect, organize, and share content (pins in Pinterest), has gained great attention in recent years. Despite the increasing interest in Pinterest, little research has paid attention to how people collect, manage, and share pins in Pinterest. In this paper, to shed insight on such issues, we study the following questions. How do people collect and manage pins by their tastes in Pinterest? What factors do mainly drive people to share their pins in Pinterest? How do the characteristics of users (e.g., gender, popularity, country) or properties of pins (e.g., category, topic) play roles in propagating pins in Pinterest? To answer these questions, we have conducted a measurement study on patterns of pin curating and sharing in Pinterest. By keeping track of all the newly posted and shared pins in each category (e.g., animal, kids, women's fashion) from June 5 to July 18, 2013, we built 350 K pin propagation trees for 3 M users. With the dataset, we investigate: (1) how users collect and curate pins, (2) how users share their pins and why, and (3) how users are related by shared pins of interest. Our key finding is that pin propagation in Pinterest is mostly driven by pin's properties like its topic, not by user's characteristics like her number of followers. We further show that users in the same community in the interest graph (i.e., representing the relations among users) of Pinterest share pins (i) in the same category with 94% probability and (ii) of the same URL where pins come from with 89% probability. Finally, we explore the implications of our findings for predicting how pins are shared in Pinterest.","PeriodicalId":306456,"journal":{"name":"Measurement and Modeling of Computer Systems","volume":"52 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2014-06-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115152802","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Switched queueing networks model wireless networks, input queued switches and numerous other networked communications systems. For single-hop networks, we consider a (α,g)-switch policy} which combines the MaxWeight policies with bandwidth sharing networks -- a further well studied model of Internet congestion. We prove the maximum stability property for this class of randomized policies. Thus these policies have the same first order behavior as the MaxWeight policies. However, for multihop networks some of these generalized polices address a number of critical weakness of the MaxWeight/BackPressure policies. For multihop networks with fixed routing, we consider the Proportional Scheduler (or (1,log)-policy). In this setting, the BackPressure policy is maximum stable, but must maintain a queue for every route-destination, which typically grows rapidly with a network's size. However, this proportionally fair policy only needs to maintain a queue for each outgoing link, which is typically bounded in number. As is common with Internet routing, by maintaining per-link queueing each node only needs to know the next hop for each packet and not its entire route. Further, in contrast to BackPressure, the Proportional Scheduler does not compare downstream queue lengths to determine weights, only local link information is required. This leads to greater potential for decomposed implementations of the policy. Through a reduction argument and an entropy argument, we demonstrate that, whilst maintaining substantially less queueing overhead, the Proportional Scheduler achieves maximum throughput stability.
{"title":"Concave switching in single and multihop networks","authors":"N. Walton","doi":"10.1145/2591971.2591987","DOIUrl":"https://doi.org/10.1145/2591971.2591987","url":null,"abstract":"Switched queueing networks model wireless networks, input queued switches and numerous other networked communications systems. For single-hop networks, we consider a (α,g)-switch policy} which combines the MaxWeight policies with bandwidth sharing networks -- a further well studied model of Internet congestion. We prove the maximum stability property for this class of randomized policies. Thus these policies have the same first order behavior as the MaxWeight policies. However, for multihop networks some of these generalized polices address a number of critical weakness of the MaxWeight/BackPressure policies.\u0000 For multihop networks with fixed routing, we consider the Proportional Scheduler (or (1,log)-policy). In this setting, the BackPressure policy is maximum stable, but must maintain a queue for every route-destination, which typically grows rapidly with a network's size. However, this proportionally fair policy only needs to maintain a queue for each outgoing link, which is typically bounded in number. As is common with Internet routing, by maintaining per-link queueing each node only needs to know the next hop for each packet and not its entire route. Further, in contrast to BackPressure, the Proportional Scheduler does not compare downstream queue lengths to determine weights, only local link information is required. This leads to greater potential for decomposed implementations of the policy. Through a reduction argument and an entropy argument, we demonstrate that, whilst maintaining substantially less queueing overhead, the Proportional Scheduler achieves maximum throughput stability.","PeriodicalId":306456,"journal":{"name":"Measurement and Modeling of Computer Systems","volume":"211 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2014-04-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134304789","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Lin Ai, X. Wu, Lingxiao Huang, Longbo Huang, Pingzhong Tang, J. Li
We consider the multi-shop ski rental problem. This problem generalizes the classic ski rental problem to a multi-shop setting, in which each shop has different prices for renting and purchasing a pair of skis, and a consumer has to make decisions on when and where to buy. We are interested in the optimal online (competitive-ratio minimizing) mixed strategy from the consumer's perspective. For our problem in its basic form, we obtain exciting closed-form solutions and a linear time algorithm for computing them. We further demonstrate the generality of our approach by investigating three extensions of our basic problem, namely ones that consider costs incurred by entering a shop or switching to another shop. Our solutions to these problems suggest that the consumer must assign positive probability in exactly one shop at any buying time. Our results apply to many real-world applications, ranging from cost management in IaaS cloud to scheduling in distributed computing.
{"title":"The multi-shop ski rental problem","authors":"Lin Ai, X. Wu, Lingxiao Huang, Longbo Huang, Pingzhong Tang, J. Li","doi":"10.1145/2591971.2591984","DOIUrl":"https://doi.org/10.1145/2591971.2591984","url":null,"abstract":"We consider the multi-shop ski rental problem. This problem generalizes the classic ski rental problem to a multi-shop setting, in which each shop has different prices for renting and purchasing a pair of skis, and a consumer has to make decisions on when and where to buy. We are interested in the optimal online (competitive-ratio minimizing) mixed strategy from the consumer's perspective. For our problem in its basic form, we obtain exciting closed-form solutions and a linear time algorithm for computing them. We further demonstrate the generality of our approach by investigating three extensions of our basic problem, namely ones that consider costs incurred by entering a shop or switching to another shop. Our solutions to these problems suggest that the consumer must assign positive probability in exactly one shop at any buying time. Our results apply to many real-world applications, ranging from cost management in IaaS cloud to scheduling in distributed computing.","PeriodicalId":306456,"journal":{"name":"Measurement and Modeling of Computer Systems","volume":"89 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2014-04-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116033607","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
In this paper, we investigate the power of online learning in stochastic network optimization with unknown system statistics a priori. We are interested in understanding how information and learning can be efficiently incorporated into system control techniques, and what are the fundamental benefits of doing so. We propose two Online Learning-Aided Control techniques, OLAC and OLAC2, that explicitly utilize the past system information in current system control via a learning procedure called dual learning. We prove strong performance guarantees of the proposed algorithms: OLAC and OLAC2 achieve the near-optimal [O(ε), O([log(1/ε)]2)] utility-delay tradeoff and OLAC2 possesses an O(ε-2/3) convergence time. Simulation results also confirm the superior performance of the proposed algorithms in practice. To the best of our knowledge, OLAC and OLAC2 are the first algorithms that simultaneously possess explicit near-optimal delay guarantee and sub-linear convergence time, and our attempt is the first to explicitly incorporate online learning into stochastic network optimization and to demonstrate its power in both theory and practice.
{"title":"The power of online learning in stochastic network optimization","authors":"Longbo Huang, Xin Liu, Xiaohong Hao","doi":"10.1145/2591971.2591990","DOIUrl":"https://doi.org/10.1145/2591971.2591990","url":null,"abstract":"In this paper, we investigate the power of online learning in stochastic network optimization with unknown system statistics <i>a priori</i>. We are interested in understanding how information and learning can be efficiently incorporated into system control techniques, and what are the fundamental benefits of doing so. We propose two <i>Online Learning-Aided Control</i> techniques, <b>OLAC</b> and <b>OLAC2</b>, that explicitly utilize the past system information in current system control via a learning procedure called <i>dual learning</i>. We prove strong performance guarantees of the proposed algorithms: <b>OLAC</b> and <b>OLAC2</b> achieve the near-optimal [<i>O</i>(ε), <i>O</i>([log(1/ε)]<sup>2</sup>)] utility-delay tradeoff and <b>OLAC2</b> possesses an <i>O</i>(ε<sup>-2/3</sup>) convergence time. Simulation results also confirm the superior performance of the proposed algorithms in practice. To the best of our knowledge, <b>OLAC</b> and <b>OLAC2</b> are the first algorithms that simultaneously possess explicit near-optimal delay guarantee and sub-linear convergence time, and our attempt is the first to explicitly incorporate online learning into stochastic network optimization and to demonstrate its power in both theory and practice.","PeriodicalId":306456,"journal":{"name":"Measurement and Modeling of Computer Systems","volume":"28 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2014-04-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127830846","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Stratis Ioannidis, A. Montanari, Udi Weinsberg, Smriti Bhagat, N. Fawaz, N. Taft
Online services routinely mine user data to predict user preferences, make recommendations, and place targeted ads. Recent research has demonstrated that several private user attributes (such as political affiliation, sexual orientation, and gender) can be inferred from such data. Can a privacy-conscious user benefit from personalization while simultaneously protecting her private attributes? We study this question in the context of a rating prediction service based on matrix factorization. We construct a protocol of interactions between the service and users that has remarkable optimality properties: it is privacy-preserving, in that no inference algorithm can succeed in inferring a user's private attribute with a probability better than random guessing; it has maximal accuracy, in that no other privacy-preserving protocol improves rating prediction; and, finally, it involves a minimal disclosure, as the prediction accuracy strictly decreases when the service reveals less information. We extensively evaluate our protocol using several rating datasets, demonstrating that it successfully blocks the inference of gender, age and political affiliation, while incurring less than 5% decrease in the accuracy of rating prediction.
{"title":"Privacy tradeoffs in predictive analytics","authors":"Stratis Ioannidis, A. Montanari, Udi Weinsberg, Smriti Bhagat, N. Fawaz, N. Taft","doi":"10.1145/2591971.2592011","DOIUrl":"https://doi.org/10.1145/2591971.2592011","url":null,"abstract":"Online services routinely mine user data to predict user preferences, make recommendations, and place targeted ads. Recent research has demonstrated that several private user attributes (such as political affiliation, sexual orientation, and gender) can be inferred from such data. Can a privacy-conscious user benefit from personalization while simultaneously protecting her private attributes? We study this question in the context of a rating prediction service based on matrix factorization. We construct a protocol of interactions between the service and users that has remarkable optimality properties: it is privacy-preserving, in that no inference algorithm can succeed in inferring a user's private attribute with a probability better than random guessing; it has maximal accuracy, in that no other privacy-preserving protocol improves rating prediction; and, finally, it involves a minimal disclosure, as the prediction accuracy strictly decreases when the service reveals less information. We extensively evaluate our protocol using several rating datasets, demonstrating that it successfully blocks the inference of gender, age and political affiliation, while incurring less than 5% decrease in the accuracy of rating prediction.","PeriodicalId":306456,"journal":{"name":"Measurement and Modeling of Computer Systems","volume":"38 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2014-03-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123540192","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Daniel S. Berger, Philipp Gland, Sahil Singla, F. Ciucu
TTL caching models have recently regained significant research interest, largely due to their ability to fit popular caching policies such as LRU. In this extended abstract we briefly describe our recent work on two exact methods to analyze TTL cache networks. The first method generalizes existing results for line networks under renewal requests to the broad class of caching policies whereby evictions are driven by stopping times. The obtained results are further generalized, using the second method, to feedforward networks with Markov arrival processes (MAP) requests. MAPs are particularly suitable for non-line networks because they are closed not only under superposition and splitting, as known, but also under input-output caching operations as proven herein for phase-type TTL distributions. The crucial benefit of the two closure properties is that they jointly enable the first exact analysis of feedforward networks of TTL caches in great generality.
{"title":"Exact analysis of TTL cache networks: the case of caching policies driven by stopping times","authors":"Daniel S. Berger, Philipp Gland, Sahil Singla, F. Ciucu","doi":"10.1145/2591971.2592038","DOIUrl":"https://doi.org/10.1145/2591971.2592038","url":null,"abstract":"TTL caching models have recently regained significant research interest, largely due to their ability to fit popular caching policies such as LRU. In this extended abstract we briefly describe our recent work on two exact methods to analyze TTL cache networks. The first method generalizes existing results for line networks under renewal requests to the broad class of caching policies whereby evictions are driven by stopping times. The obtained results are further generalized, using the second method, to feedforward networks with Markov arrival processes (MAP) requests. MAPs are particularly suitable for non-line networks because they are closed not only under superposition and splitting, as known, but also under input-output caching operations as proven herein for phase-type TTL distributions. The crucial benefit of the two closure properties is that they jointly enable the first exact analysis of feedforward networks of TTL caches in great generality.","PeriodicalId":306456,"journal":{"name":"Measurement and Modeling of Computer Systems","volume":"19 19","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2014-02-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"113941727","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Jiaming Xu, Rui Wu, Kai Zhu, B. Hajek, R. Srikant, Lei Ying
In standard clustering problems, data points are represented by vectors, and by stacking them together, one forms a data matrix with row or column cluster structure. In this paper, we consider a class of binary matrices, arising in many applications, which exhibit both row and column cluster structure, and our goal is to exactly recover the underlying row and column clusters by observing only a small fraction of noisy entries. We first derive a lower bound on the minimum number of observations needed for exact cluster recovery. Then, we study three algorithms with different running time and compare the number of observations needed by them for successful cluster recovery. Our analytical results show smooth time-data trade offs: one can gradually reduce the computational complexity when increasingly more observations are available.
{"title":"Jointly clustering rows and columns of binary matrices: algorithms and trade-offs","authors":"Jiaming Xu, Rui Wu, Kai Zhu, B. Hajek, R. Srikant, Lei Ying","doi":"10.1145/2591971.2592005","DOIUrl":"https://doi.org/10.1145/2591971.2592005","url":null,"abstract":"In standard clustering problems, data points are represented by vectors, and by stacking them together, one forms a data matrix with row or column cluster structure. In this paper, we consider a class of binary matrices, arising in many applications, which exhibit both row and column cluster structure, and our goal is to exactly recover the underlying row and column clusters by observing only a small fraction of noisy entries. We first derive a lower bound on the minimum number of observations needed for exact cluster recovery. Then, we study three algorithms with different running time and compare the number of observations needed by them for successful cluster recovery. Our analytical results show smooth time-data trade offs: one can gradually reduce the computational complexity when increasingly more observations are available.","PeriodicalId":306456,"journal":{"name":"Measurement and Modeling of Computer Systems","volume":"22 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2013-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116762105","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Current public cloud offerings sell capacity in the form of pre-defined virtual machine (VM) configurations to their tenants. Typically this means that tenants must purchase individual VM configurations based on the peak demands of the applications, or be restricted to only scale-out applications that can share a pool of VMs. This diminishes the value proposition of moving to a public cloud as compared to server consolidation in a private virtualized datacenter, where one gets the benefits of statistical multiplexing between VMs belonging to the same or different applications. Ideally one would like to enable a cloud tenant to buy capacity in bulk and benefit from statistical multiplexing among its workloads. This requires the purchased capacity to be dynamically and transparently allocated among the tenant's VMs that may be running on different servers, even across datacenters. In this paper, we propose two novel algorithms called BPX and DBS that are able to provide the cloud customer with the abstraction of buying bulk capacity. These algorithms dynamically allocate the bulk capacity purchased by a customer between its VMs based on their individual demands and user-set importance. Our algorithms are highly scalable and are designed to work in a large-scale distributed environment. We implemented a prototype of BPX as part of VMware's management software and showed that BPX is able to closely mimic the behavior of a centralized allocator in a distributed manner.
{"title":"Defragmenting the cloud using demand-based resource allocation","authors":"Ganesha Shanmuganathan, Ajay Gulati, P. Varman","doi":"10.1145/2465529.2465763","DOIUrl":"https://doi.org/10.1145/2465529.2465763","url":null,"abstract":"Current public cloud offerings sell capacity in the form of pre-defined virtual machine (VM) configurations to their tenants. Typically this means that tenants must purchase individual VM configurations based on the peak demands of the applications, or be restricted to only scale-out applications that can share a pool of VMs. This diminishes the value proposition of moving to a public cloud as compared to server consolidation in a private virtualized datacenter, where one gets the benefits of statistical multiplexing between VMs belonging to the same or different applications. Ideally one would like to enable a cloud tenant to buy capacity in bulk and benefit from statistical multiplexing among its workloads. This requires the purchased capacity to be dynamically and transparently allocated among the tenant's VMs that may be running on different servers, even across datacenters. In this paper, we propose two novel algorithms called BPX and DBS that are able to provide the cloud customer with the abstraction of buying bulk capacity. These algorithms dynamically allocate the bulk capacity purchased by a customer between its VMs based on their individual demands and user-set importance. Our algorithms are highly scalable and are designed to work in a large-scale distributed environment. We implemented a prototype of BPX as part of VMware's management software and showed that BPX is able to closely mimic the behavior of a centralized allocator in a distributed manner.","PeriodicalId":306456,"journal":{"name":"Measurement and Modeling of Computer Systems","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2013-06-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122618720","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Ming Li, Andrey Lukyanenko, S. Tarkoma, Yong Cui, Antti Ylä-Jääski
Abstract With bounded receive buffers, the aggregate bandwidth of multipath transmission degrades significantly in the presence of path heterogeneity. The performance could even be worse than that of single-path TCP, undermining the advantage gained by using multipath transmit. Furthermore, multipath transmission also suffers from delay and jitter even with large receive buffers. In order to tolerate the path heterogeneity when the receive buffer is bounded, we propose a new multipath TCP protocol, namely SC-MPTCP, by integrating linear systematic coding into MPTCP. In SC-MPTCP, we make use of coded packets as redundancy to counter against expensive retransmissions. The redundancy is provisioned into both proactive and reactive data. Specifically, to send a generation of packets, SC-MPTCP transmits proactive redundancy first and then delivers the original packets, instead of encoding all sent-out packets as all the existing coding solutions have done. The proactive redundancy is continuously updated according to the estimated aggregate retransmission ratio. In order to avoid the proactive redundancy being underestimated, the pre-blocking warning mechanism is utilized to retrieve the reactive redundancy from the sender. We use an NS-3 network simulator to evaluate the performance of SC-MPTCP with and without the coupled congestion control option. The results show that with bounded receive buffers, MPTCP achieves less than 20% of the optimal goodput with diverse packet losses, whereas SC-MPTCP approaches the optimal performance with significantly smaller receive buffers. With the help of systematic coding, SC-MPTCP reduces the average buffer delay of MPTCP by at least 80% in different test scenarios. We also demonstrate that the use of systematic coding could significantly reduce the arithmetic complexity compared with the use of non-systematic coding.
{"title":"Tolerating path heterogeneity in multipath TCP with bounded receive buffers","authors":"Ming Li, Andrey Lukyanenko, S. Tarkoma, Yong Cui, Antti Ylä-Jääski","doi":"10.1145/2465529.2465750","DOIUrl":"https://doi.org/10.1145/2465529.2465750","url":null,"abstract":"Abstract With bounded receive buffers, the aggregate bandwidth of multipath transmission degrades significantly in the presence of path heterogeneity. The performance could even be worse than that of single-path TCP, undermining the advantage gained by using multipath transmit. Furthermore, multipath transmission also suffers from delay and jitter even with large receive buffers. In order to tolerate the path heterogeneity when the receive buffer is bounded, we propose a new multipath TCP protocol, namely SC-MPTCP, by integrating linear systematic coding into MPTCP. In SC-MPTCP, we make use of coded packets as redundancy to counter against expensive retransmissions. The redundancy is provisioned into both proactive and reactive data. Specifically, to send a generation of packets, SC-MPTCP transmits proactive redundancy first and then delivers the original packets, instead of encoding all sent-out packets as all the existing coding solutions have done. The proactive redundancy is continuously updated according to the estimated aggregate retransmission ratio. In order to avoid the proactive redundancy being underestimated, the pre-blocking warning mechanism is utilized to retrieve the reactive redundancy from the sender. We use an NS-3 network simulator to evaluate the performance of SC-MPTCP with and without the coupled congestion control option. The results show that with bounded receive buffers, MPTCP achieves less than 20% of the optimal goodput with diverse packet losses, whereas SC-MPTCP approaches the optimal performance with significantly smaller receive buffers. With the help of systematic coding, SC-MPTCP reduces the average buffer delay of MPTCP by at least 80% in different test scenarios. We also demonstrate that the use of systematic coding could significantly reduce the arithmetic complexity compared with the use of non-systematic coding.","PeriodicalId":306456,"journal":{"name":"Measurement and Modeling of Computer Systems","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2013-06-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134216950","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：