Performance and reliability are two core challenges for today's cloud data centers. Emerging non-volatile memory (NVM) technologies, which promise large capacity, high-speed, byte-addressable and persistent memory, can offer mitigating benefits. In particular, business-critical applications in ecommerce, finance, and banking could persist transactions in NVM either as traditional storage or directly as durable memory, after additional development to adapt the applications to the new interface. However, both approaches have very diverse software overheads that in the literature are still scantly compared head-to-head or clearly quantified. In order to shed some light on these issues, we developed a suite of throughput, latency, and scalability tests that focus on the challenge of persisting financial transactions in the form of small and critical parcels of data, a representative challenge for financial cloud data centers. By carrying out benchmarks on a real NVDIMM server, we compare and contrast in detail the performance of the programming framework Mnemosyne with the NVM storage solutions PMFS (a persistent memory file system) and PMBD (a persistent memory block-device). In turn, these are compared with both directly-addressable volatile RAM and a fast NVM Express flash drive (NVMe) as performance baselines. We found that persisting financial transactions with Mnemosyne achieves up to two orders of magnitude better throughput than persisting them in the NVMe, while incurring a performance penalty of 25 percent over volatile RAM. Furthermore, committing transactions in NVM as persistent memory or flat files is up to two orders of magnitude faster than persisting them in databases saved in NVM. Finally, the throughput of writing financial transactions using Mnemosyne is four times higher than PMFS and one order of magnitude higher than PMBD.
{"title":"Efficient Persistence of Financial Transactions in NVM-based Cloud Data Centers","authors":"S. Ruocco, Duy-Khanh Le","doi":"10.1109/ICCCRI.2015.29","DOIUrl":"https://doi.org/10.1109/ICCCRI.2015.29","url":null,"abstract":"Performance and reliability are two core challenges for today's cloud data centers. Emerging non-volatile memory (NVM) technologies, which promise large capacity, high-speed, byte-addressable and persistent memory, can offer mitigating benefits. In particular, business-critical applications in ecommerce, finance, and banking could persist transactions in NVM either as traditional storage or directly as durable memory, after additional development to adapt the applications to the new interface. However, both approaches have very diverse software overheads that in the literature are still scantly compared head-to-head or clearly quantified. In order to shed some light on these issues, we developed a suite of throughput, latency, and scalability tests that focus on the challenge of persisting financial transactions in the form of small and critical parcels of data, a representative challenge for financial cloud data centers. By carrying out benchmarks on a real NVDIMM server, we compare and contrast in detail the performance of the programming framework Mnemosyne with the NVM storage solutions PMFS (a persistent memory file system) and PMBD (a persistent memory block-device). In turn, these are compared with both directly-addressable volatile RAM and a fast NVM Express flash drive (NVMe) as performance baselines. We found that persisting financial transactions with Mnemosyne achieves up to two orders of magnitude better throughput than persisting them in the NVMe, while incurring a performance penalty of 25 percent over volatile RAM. Furthermore, committing transactions in NVM as persistent memory or flat files is up to two orders of magnitude faster than persisting them in databases saved in NVM. Finally, the throughput of writing financial transactions using Mnemosyne is four times higher than PMFS and one order of magnitude higher than PMBD.","PeriodicalId":183970,"journal":{"name":"2015 International Conference on Cloud Computing Research and Innovation (ICCCRI)","volume":"46 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2015-10-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125824200","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}
Next generation clouds, based on highly programmable, high performance networks, especially those supported by Software-Defined-Networking (SDN) have attracted significant interest by research communities. In recognition of the increasing importance of advancing cloud services and technologies, especially for providing Internet services, the US National Science Foundation (NSF) established a project, the NSF Cloud initiative, to enable the computer science research community to develop and experiment with novel cloud architectures and create new, architecturally enabled innovative applications for cloud computing through empirical research experimentation by using large scale distributed cloud test beds. This paper provides an overview of one of those test beds, the Chameleon Cloud tested, with an additional description of the integration of that test bed with high programmable, high performance networks, based on SDN. The Chameleon project is designing, deploying, and operating a large scale, highly distributed experimental environment for empirical cloud research, integrated with high programmable networks as a foundation resource.
{"title":"Next Generation Clouds, the Chameleon Cloud Testbed, and Software Defined Networking (SDN)","authors":"J. Mambretti, J. Chen, F. Yeh","doi":"10.1109/ICCCRI.2015.10","DOIUrl":"https://doi.org/10.1109/ICCCRI.2015.10","url":null,"abstract":"Next generation clouds, based on highly programmable, high performance networks, especially those supported by Software-Defined-Networking (SDN) have attracted significant interest by research communities. In recognition of the increasing importance of advancing cloud services and technologies, especially for providing Internet services, the US National Science Foundation (NSF) established a project, the NSF Cloud initiative, to enable the computer science research community to develop and experiment with novel cloud architectures and create new, architecturally enabled innovative applications for cloud computing through empirical research experimentation by using large scale distributed cloud test beds. This paper provides an overview of one of those test beds, the Chameleon Cloud tested, with an additional description of the integration of that test bed with high programmable, high performance networks, based on SDN. The Chameleon project is designing, deploying, and operating a large scale, highly distributed experimental environment for empirical cloud research, integrated with high programmable networks as a foundation resource.","PeriodicalId":183970,"journal":{"name":"2015 International Conference on Cloud Computing Research and Innovation (ICCCRI)","volume":"156 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2015-10-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127357792","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}
Fast evolving smart phone technology greatly promoted consumption of service and content on the move, meanwhile raised new privacy, security and health issues, e.g., Increasing Pedestrians Multitasking with Smart phones (PMS) fall injury. Existing camera or vehicular-based safety technologies mostly aims to warn PMS based on camera observations which are limited to the coverage of camera view. To enable warning without camera, we look to study PMS injury issue by understand the internal causes of the increased injury, say the distractive multitasking phone usage activities (MPUA). In this paper, we present Safe MT: a smart phone-based PMS safety application to case study the suspicious typical MPUA that increase fall risk in daily life. Safe MT provides accurate private monitoring of phone usage activity (PUA) as well as accompanying gait style (GS). For PUA monitoring, Safe MT employees a novel out-of-band approach to infer typical PUA, e.g., Calling, messaging, without causing privacy harshness at low overhead. For efficient GS monitoring, Safe MT employed a novel gait-style classification (GSC) algorithm that overcome the challenges of subjective gait style signature, using later-binding initialization with subjective user data. We implemented the system on Android phones and validated its availability in supervised lab experiments, results show that our system can effectively identify the two parameters of MPUA. Although fall injury case are hardly recorded in 4-week real trace data, understanding of MPUA are still gained by mining the collected dataset.
{"title":"Towards a Private Fall Injury Warning Service for Smartphone-Distracted Pedestrian","authors":"Jianxiong Yin, Yonggang Wen, Jianxin Wu","doi":"10.1109/ICCCRI.2015.15","DOIUrl":"https://doi.org/10.1109/ICCCRI.2015.15","url":null,"abstract":"Fast evolving smart phone technology greatly promoted consumption of service and content on the move, meanwhile raised new privacy, security and health issues, e.g., Increasing Pedestrians Multitasking with Smart phones (PMS) fall injury. Existing camera or vehicular-based safety technologies mostly aims to warn PMS based on camera observations which are limited to the coverage of camera view. To enable warning without camera, we look to study PMS injury issue by understand the internal causes of the increased injury, say the distractive multitasking phone usage activities (MPUA). In this paper, we present Safe MT: a smart phone-based PMS safety application to case study the suspicious typical MPUA that increase fall risk in daily life. Safe MT provides accurate private monitoring of phone usage activity (PUA) as well as accompanying gait style (GS). For PUA monitoring, Safe MT employees a novel out-of-band approach to infer typical PUA, e.g., Calling, messaging, without causing privacy harshness at low overhead. For efficient GS monitoring, Safe MT employed a novel gait-style classification (GSC) algorithm that overcome the challenges of subjective gait style signature, using later-binding initialization with subjective user data. We implemented the system on Android phones and validated its availability in supervised lab experiments, results show that our system can effectively identify the two parameters of MPUA. Although fall injury case are hardly recorded in 4-week real trace data, understanding of MPUA are still gained by mining the collected dataset.","PeriodicalId":183970,"journal":{"name":"2015 International Conference on Cloud Computing Research and Innovation (ICCCRI)","volume":"135 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2015-10-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122174673","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}
The cloud storage services bring forth infinite storage capacity and flexible access capability to store and share large-scale content. The convenience brought forth has attracted both individual and enterprise users to outsource data service to a cloud provider. As the survey shows 56% of the usages of cloud storage applications are for data back up and up to 68% of data backup are user assets. Enterprise tenants would need to protect their data privacy before uploading them to the cloud and expect a reasonable performance while they try to reduce the operation cost in terms of cloud storage, capacity and I/Os matter as well as systems' performance, bandwidth and data protection. Thus, enterprise tenants demand secure and economic data storage yet flexible access on their cloud data. In this paper, we propose a secure de-duplication solution for enterprise tenants to leverage the benefits of cloud storage while reducing operation cost and protecting privacy. First, the solution uses a proxy to do flexible group access control which supports secure de-duplication within a group, Second, the solution supports scalable clustering of proxies to support large-scale data access, Third, the solution can be integrated with cloud storage seamlessly. We implemented and tested our solution by integrating it with Drop box. Secure de-duplication in a group is performed at low data transfer latency and small storage overhead as compared to de-duplication on plaintext.
{"title":"Flexible Yet Secure De-duplication Service for Enterprise Data on Cloud Storage","authors":"W. Chuan, Shu Qin Ren, S. Keoh, Khin Mi Mi Aung","doi":"10.1109/ICCCRI.2015.11","DOIUrl":"https://doi.org/10.1109/ICCCRI.2015.11","url":null,"abstract":"The cloud storage services bring forth infinite storage capacity and flexible access capability to store and share large-scale content. The convenience brought forth has attracted both individual and enterprise users to outsource data service to a cloud provider. As the survey shows 56% of the usages of cloud storage applications are for data back up and up to 68% of data backup are user assets. Enterprise tenants would need to protect their data privacy before uploading them to the cloud and expect a reasonable performance while they try to reduce the operation cost in terms of cloud storage, capacity and I/Os matter as well as systems' performance, bandwidth and data protection. Thus, enterprise tenants demand secure and economic data storage yet flexible access on their cloud data. In this paper, we propose a secure de-duplication solution for enterprise tenants to leverage the benefits of cloud storage while reducing operation cost and protecting privacy. First, the solution uses a proxy to do flexible group access control which supports secure de-duplication within a group, Second, the solution supports scalable clustering of proxies to support large-scale data access, Third, the solution can be integrated with cloud storage seamlessly. We implemented and tested our solution by integrating it with Drop box. Secure de-duplication in a group is performed at low data transfer latency and small storage overhead as compared to de-duplication on plaintext.","PeriodicalId":183970,"journal":{"name":"2015 International Conference on Cloud Computing Research and Innovation (ICCCRI)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2015-10-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132964240","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}
Hadoop is an open-source software framework for distributed computing that is widely used to develop large-scale data processing applications, such as big data applications. Hadoop application programs are normally run on in-house or cloud computing platforms. Recently, a hybrid cloud composed of in-house and remote cloud computing platforms has been found to be capable of sustaining a certain level of application performance. In this paper, we discuss the performance of a Hadoop application program running on such hybrid clouds. We will begin by presenting the performance model used to estimate the execution time of a Hadoop application program running on a hybrid cloud. Then, we will show the results of experiments conducted on hybrid cloud test beds. These experimental results revealed that the performance levels of the Hadoop application programs running on the hybrid cloud were application type dependent, and that performance improvements could be expected by using a remote cloud computing platform in conjunction with in-house computing platforms for certain types of applications. Furthermore, the results showed that our performance model captured the performance trend of the application programs on the hybrid cloud. However, room for improvement still exists in the performance model, particularly for the shuffle phase.
{"title":"Performance of Hadoop Application on Hybrid Cloud","authors":"H. Ohnaga, K. Aida, Omar Abdul-Rahman","doi":"10.1109/ICCCRI.2015.25","DOIUrl":"https://doi.org/10.1109/ICCCRI.2015.25","url":null,"abstract":"Hadoop is an open-source software framework for distributed computing that is widely used to develop large-scale data processing applications, such as big data applications. Hadoop application programs are normally run on in-house or cloud computing platforms. Recently, a hybrid cloud composed of in-house and remote cloud computing platforms has been found to be capable of sustaining a certain level of application performance. In this paper, we discuss the performance of a Hadoop application program running on such hybrid clouds. We will begin by presenting the performance model used to estimate the execution time of a Hadoop application program running on a hybrid cloud. Then, we will show the results of experiments conducted on hybrid cloud test beds. These experimental results revealed that the performance levels of the Hadoop application programs running on the hybrid cloud were application type dependent, and that performance improvements could be expected by using a remote cloud computing platform in conjunction with in-house computing platforms for certain types of applications. Furthermore, the results showed that our performance model captured the performance trend of the application programs on the hybrid cloud. However, room for improvement still exists in the performance model, particularly for the shuffle phase.","PeriodicalId":183970,"journal":{"name":"2015 International Conference on Cloud Computing Research and Innovation (ICCCRI)","volume":"59 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2015-10-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130133503","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}
Cloud Computing may be viewed as a utility-based service, similar to the use of (e.g.) Electricity. The setting up of the service, management thereof, support and subsequent upgrades are performed by the CSP (Cloud Service Provider) on behalf of the user. The user subscribes to the service and pays for the computing resources that were used during the specified period. Service subscription may be across the whole technology stack from Applications, Infrastructure and Platform, and with options of how it should be deployed i.e. Privately (in-house), Community (for limited number of users), Public (consumed via internet with no dedicated firewalls) or hybrid (with both in-house and public deployments). In this paper the researchers argue that Cloud Computing can alleviate the negative impact of poor management, lack of skill, lack of funding, etc. On the success of a small or medium business. Following a literature survey on the challenges experienced by Small and Medium-sized Enterprises (SMEs), with specific focus on the most recent South African SBP, SME Growth index, 2013, we suggest how Cloud Computing may facilitate SME business growth.
{"title":"The Role of Cloud Computing in Addressing SME Challenges in South Africa","authors":"Nkosi Kumalo, J. A. V. D. Poll","doi":"10.1109/ICCCRI.2015.32","DOIUrl":"https://doi.org/10.1109/ICCCRI.2015.32","url":null,"abstract":"Cloud Computing may be viewed as a utility-based service, similar to the use of (e.g.) Electricity. The setting up of the service, management thereof, support and subsequent upgrades are performed by the CSP (Cloud Service Provider) on behalf of the user. The user subscribes to the service and pays for the computing resources that were used during the specified period. Service subscription may be across the whole technology stack from Applications, Infrastructure and Platform, and with options of how it should be deployed i.e. Privately (in-house), Community (for limited number of users), Public (consumed via internet with no dedicated firewalls) or hybrid (with both in-house and public deployments). In this paper the researchers argue that Cloud Computing can alleviate the negative impact of poor management, lack of skill, lack of funding, etc. On the success of a small or medium business. Following a literature survey on the challenges experienced by Small and Medium-sized Enterprises (SMEs), with specific focus on the most recent South African SBP, SME Growth index, 2013, we suggest how Cloud Computing may facilitate SME business growth.","PeriodicalId":183970,"journal":{"name":"2015 International Conference on Cloud Computing Research and Innovation (ICCCRI)","volume":"94 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2015-10-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121840309","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}
Miguel Rodel Felipe, Khin Mi Mi Aung, Xia Ye, Yonggang Wen
Customer Relationship Management (CRM) system improves companies' profitability by helping companies focus on the relationships with customers, colleagues or suppliers. By having strong initiative to move applications to cloud, enterprises are hindered by cloud security and reliability issues [1], especially when it comes to financial industries. To provide a practical and secure solution to these enterprises, this project aims to build a cloud CRM system that enables fully homomorphic encryption. In order to explore the potential of this, the project integrates three key components: Open source CRM system Sugar CRM, partial homomorphic database system Crypt DB and fully homomorphic encryption library HElib. By leveraging the structure based on our previous work [2], Stealthy CRM successfully integrates fully homomorphic encryption support on top of Crypt DB database encryption environment. Besides that, Stealthy CRM enables a transparent and seamless integration to any CRM system by using a modified My SQL proxy to listen to, encrypt the queries and interact with Crypt DB and HElib subsystems. An evaluation of TPC-C and TPC-H queries is conducted on Stealthy CRM system. The result shows Stealthy CRM has 14%-28% throughput overhead for most of the CRM queries, compared with unmodified My SQL server. For complex TPC-H queries involving multiplication and composition of computation, Stealthy CRM is able to execute the query between 1.75 min to 11.7 min. Although the time takes to complete a fully homomorphic query in CRM system is still long, Stealthy CRM provided a prototype for researchers and other business application developers to explore the potential.
{"title":"StealthyCRM: A Secure Cloud CRM System Application that Supports Fully Homomorphic Database Encryption","authors":"Miguel Rodel Felipe, Khin Mi Mi Aung, Xia Ye, Yonggang Wen","doi":"10.1109/ICCCRI.2015.23","DOIUrl":"https://doi.org/10.1109/ICCCRI.2015.23","url":null,"abstract":"Customer Relationship Management (CRM) system improves companies' profitability by helping companies focus on the relationships with customers, colleagues or suppliers. By having strong initiative to move applications to cloud, enterprises are hindered by cloud security and reliability issues [1], especially when it comes to financial industries. To provide a practical and secure solution to these enterprises, this project aims to build a cloud CRM system that enables fully homomorphic encryption. In order to explore the potential of this, the project integrates three key components: Open source CRM system Sugar CRM, partial homomorphic database system Crypt DB and fully homomorphic encryption library HElib. By leveraging the structure based on our previous work [2], Stealthy CRM successfully integrates fully homomorphic encryption support on top of Crypt DB database encryption environment. Besides that, Stealthy CRM enables a transparent and seamless integration to any CRM system by using a modified My SQL proxy to listen to, encrypt the queries and interact with Crypt DB and HElib subsystems. An evaluation of TPC-C and TPC-H queries is conducted on Stealthy CRM system. The result shows Stealthy CRM has 14%-28% throughput overhead for most of the CRM queries, compared with unmodified My SQL server. For complex TPC-H queries involving multiplication and composition of computation, Stealthy CRM is able to execute the query between 1.75 min to 11.7 min. Although the time takes to complete a fully homomorphic query in CRM system is still long, Stealthy CRM provided a prototype for researchers and other business application developers to explore the potential.","PeriodicalId":183970,"journal":{"name":"2015 International Conference on Cloud Computing Research and Innovation (ICCCRI)","volume":"9 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2015-10-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123929178","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}
Zong-Gan Chen, Zhi-hui Zhan, Hai-Hao Li, Ke-Jing Du, J. Zhong, Y. W. Foo, Yun Li, Jun Zhang
Cloud computing resources scheduling is essential for executing workflows in the cloud platform because it relates to both execution time and execution cost. In this paper, we adopt a model that optimizes the execution cost while meeting deadline constraints. In solving this problem, we propose an Improved Ant Colony System (IACS) approach featuring two novel strategies. Firstly, a dynamic heuristic strategy is used to calculate a heuristic value during an evolutionary process by taking the workflow topological structure into consideration. Secondly, a double search strategy is used to initialize the pheromone and calculate the heuristic value according to the execution time at the beginning and to initialize the pheromone and calculate heuristic value according to the execution cost after a feasible solution is found. Therefore, the proposed IACS is adaptive to the search environment and to different objectives. We have conducted extensive experiments based on workflows with different scales and different cloud resources. We compare the result with a particle swarm optimization (PSO) approach and a dynamic objective genetic algorithm (DOGA) approach. Experimental results show that IACS is able to find better solutions with a lower cost than both PSO and DOGA do on various scheduling scales and deadline conditions.
{"title":"Deadline Constrained Cloud Computing Resources Scheduling through an Ant Colony System Approach","authors":"Zong-Gan Chen, Zhi-hui Zhan, Hai-Hao Li, Ke-Jing Du, J. Zhong, Y. W. Foo, Yun Li, Jun Zhang","doi":"10.1109/ICCCRI.2015.14","DOIUrl":"https://doi.org/10.1109/ICCCRI.2015.14","url":null,"abstract":"Cloud computing resources scheduling is essential for executing workflows in the cloud platform because it relates to both execution time and execution cost. In this paper, we adopt a model that optimizes the execution cost while meeting deadline constraints. In solving this problem, we propose an Improved Ant Colony System (IACS) approach featuring two novel strategies. Firstly, a dynamic heuristic strategy is used to calculate a heuristic value during an evolutionary process by taking the workflow topological structure into consideration. Secondly, a double search strategy is used to initialize the pheromone and calculate the heuristic value according to the execution time at the beginning and to initialize the pheromone and calculate heuristic value according to the execution cost after a feasible solution is found. Therefore, the proposed IACS is adaptive to the search environment and to different objectives. We have conducted extensive experiments based on workflows with different scales and different cloud resources. We compare the result with a particle swarm optimization (PSO) approach and a dynamic objective genetic algorithm (DOGA) approach. Experimental results show that IACS is able to find better solutions with a lower cost than both PSO and DOGA do on various scheduling scales and deadline conditions.","PeriodicalId":183970,"journal":{"name":"2015 International Conference on Cloud Computing Research and Innovation (ICCCRI)","volume":"25 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2015-10-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124724761","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}
People and things become mobile sensors that converge to our daily life. This has unwittingly collected humongous of time series of data with location. People are finding ways to turn this raw data into valuable information as a distinguished business analytic. Importantly, the demand of speedy computation with an appealing visualization is crucial to success. Thus, it reveals the potential economic benefits and becomes an overwhelming new research area that requiring sophisticated mechanisms and technologies to reach the demand. Over the past decade, there have attempts of using accelerators along with multicore CPUs in boosting large-scale data computation. We proposed an emerging SQL-like GPU query accelerator, Galactic a DB. In addition, we extended it to have the geo-spatial compute capabilities. The query operation executes parallelly with drawing support from a high performance and energy efficient NVIDIA Tesla technology. Our result has shown the significant speedup by using Galactic a DB.
人和物成为移动传感器，融入我们的日常生活。这无意中收集了大量的时间序列数据与位置。人们正在寻找方法，将这些原始数据转化为有价值的信息，作为杰出的业务分析。重要的是，对快速计算和吸引人的可视化的需求是成功的关键。因此，它揭示了潜在的经济效益，成为一个压倒性的新研究领域，需要复杂的机制和技术来达到需求。在过去的十年中，人们尝试使用加速器和多核cpu来提高大规模数据计算。我们提出了一个新兴的类似sql的GPU查询加速器，Galactic a DB。此外，我们对其进行了扩展，使其具有地理空间计算能力。查询操作在高性能和节能的NVIDIA Tesla技术的支持下并行执行。我们的结果显示了使用Galactic a DB的显著加速。
{"title":"A GPU Query Accelerator for Geospatial Coordinates Computation","authors":"K. Yong, W. K. Ho, M. Chua, S. See","doi":"10.1109/ICCCRI.2015.26","DOIUrl":"https://doi.org/10.1109/ICCCRI.2015.26","url":null,"abstract":"People and things become mobile sensors that converge to our daily life. This has unwittingly collected humongous of time series of data with location. People are finding ways to turn this raw data into valuable information as a distinguished business analytic. Importantly, the demand of speedy computation with an appealing visualization is crucial to success. Thus, it reveals the potential economic benefits and becomes an overwhelming new research area that requiring sophisticated mechanisms and technologies to reach the demand. Over the past decade, there have attempts of using accelerators along with multicore CPUs in boosting large-scale data computation. We proposed an emerging SQL-like GPU query accelerator, Galactic a DB. In addition, we extended it to have the geo-spatial compute capabilities. The query operation executes parallelly with drawing support from a high performance and energy efficient NVIDIA Tesla technology. Our result has shown the significant speedup by using Galactic a DB.","PeriodicalId":183970,"journal":{"name":"2015 International Conference on Cloud Computing Research and Innovation (ICCCRI)","volume":"6 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2015-10-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116834884","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}
Fairness and data locality are often in conflict in Hadoop job scheduling. During scheduling, it is not always possible for data locality to be achieved for all jobs or for fairness to be attained for all users. Achieving pure fairness may compromise the data locality of the jobs which will negatively affect performances, and vice-versa. For example, a scheduler may opt to sacrifice performance by scheduling tasks to non-data local nodes. Alternatively, a scheduler may choose to sacrifice fairness by giving up an available slot and wait for a data-local node. The Dynamic Task Splitting Scheduler (DTSS) is proposed to mitigate the tradeoffs between fairness and data locality during job scheduling. DTSS does so by dynamically splitting a task and executing the split task immediately, on a non-data-local node, to improve the fairness. Analysis and experiments results show that it is possible to improve both fairness and the performance by adjusting the proportion of the task split. DTSS is shown to improve the make span of different users in a cluster by 2% to 11% as compared to delay scheduling under the situation where it is difficult to obtain data-local nodes on a cluster. Lastly, experiments show that DTSS is not a suitable scheduler under conditions where jobs are able to obtain data-local nodes easily.
{"title":"Hadoop Job Scheduling with Dynamic Task Splitting","authors":"YongLiang Xu, Wentong Cai","doi":"10.1109/ICCCRI.2015.31","DOIUrl":"https://doi.org/10.1109/ICCCRI.2015.31","url":null,"abstract":"Fairness and data locality are often in conflict in Hadoop job scheduling. During scheduling, it is not always possible for data locality to be achieved for all jobs or for fairness to be attained for all users. Achieving pure fairness may compromise the data locality of the jobs which will negatively affect performances, and vice-versa. For example, a scheduler may opt to sacrifice performance by scheduling tasks to non-data local nodes. Alternatively, a scheduler may choose to sacrifice fairness by giving up an available slot and wait for a data-local node. The Dynamic Task Splitting Scheduler (DTSS) is proposed to mitigate the tradeoffs between fairness and data locality during job scheduling. DTSS does so by dynamically splitting a task and executing the split task immediately, on a non-data-local node, to improve the fairness. Analysis and experiments results show that it is possible to improve both fairness and the performance by adjusting the proportion of the task split. DTSS is shown to improve the make span of different users in a cluster by 2% to 11% as compared to delay scheduling under the situation where it is difficult to obtain data-local nodes on a cluster. Lastly, experiments show that DTSS is not a suitable scheduler under conditions where jobs are able to obtain data-local nodes easily.","PeriodicalId":183970,"journal":{"name":"2015 International Conference on Cloud Computing Research and Innovation (ICCCRI)","volume":"47 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2015-10-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132456689","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：