2016 2nd International Conference on Green High Performance Computing (ICGHPC)最新文献_第2页

Adaptive Power Management for HPC applications HPC应用的自适应电源管理

2016 2nd International Conference on Green High Performance Computing (ICGHPC) Pub Date : 2016-02-01 DOI: 10.1109/ICGHPC.2016.7508065

S. Saurav, L. GangaPrasadG., Manisha Chauhan

{"title":"Adaptive Power Management for HPC applications","authors":"S. Saurav, L. GangaPrasadG., Manisha Chauhan","doi":"10.1109/ICGHPC.2016.7508065","DOIUrl":"https://doi.org/10.1109/ICGHPC.2016.7508065","url":null,"abstract":"Reduction of power and energy consumption is one of the major concerns and challenges for High Performance Computing (HPC). However, as we move towards Exascale, it will be power limited in future. The advent of Running Average Power Limit (RAPL) abstraction after sandy bridge processors has paved the way to manage power adaptively. This gives the fine-grained power measurement and control mechanism with integrated voltage regulator at core level based on power budget. The purpose of Adaptive Power Management System (APM) for HPC systems is to decide when to place power manageable components into various power saving states based on the power consumption of an application within the power budget. In HPC, jobs are distributed across various computing nodes and power management is more complex with respect to the placement of the components in required operating states. The real time power monitoring and controlling through RAPL interface gives an opportunity for adaptive management. In this paper, we describe fine-grained profiling of HPC applications and control mechanism at component level using RAPL and APM system. The idea is to profile the HPC applications at fine granular level by measuring the power consumed by various power manageable components of a node such as processors and DRAM so that more accurate power related information can be obtained and then adaptively learn and devise the Optimal Power Budget (OPB) of an application. The OPB information is stored in Knowledge Base (KB). The devised OPB for HPC application with optimal number of processors is incorporated in the job scheduler to take power-aware scheduling decision.","PeriodicalId":268630,"journal":{"name":"2016 2nd International Conference on Green High Performance Computing (ICGHPC)","volume":"4 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2016-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123920493","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}

引用次数: 5

Node level Power Profiling and Thermal Management in HPC system 高性能计算系统中的节点级功率分析和热管理

2016 2nd International Conference on Green High Performance Computing (ICGHPC) Pub Date : 2016-02-01 DOI: 10.1109/ICGHPC.2016.7508064

A. SherinM., V. ArunKumar, P. Prasanth, R. Vasudevan, J. Shamshudeen

{"title":"Node level Power Profiling and Thermal Management in HPC system","authors":"A. SherinM., V. ArunKumar, P. Prasanth, R. Vasudevan, J. Shamshudeen","doi":"10.1109/ICGHPC.2016.7508064","DOIUrl":"https://doi.org/10.1109/ICGHPC.2016.7508064","url":null,"abstract":"In addition to the performance, power consumption has become a major concern in High Performance Computing (HPC) systems. Typically the cooling system and the IT loads are the major contributors to the power bills. Understanding the power consumption at different granular levels in the HPC system is a first step to quantify the problems in HPC system. By proper monitoring and effective utilization of the cooling system, the power requirements of the HPC facility can be effectively met. In this paper we present a system where node level power measurement and WSN based rack level temperature measurement are used to provide localized control of cold air supply. A Smart Power Monitoring and Distribution Unit (PMDU) is designed and developed, to replace the existing Power Distribution Unit (PDU) in HPC. This can measure and report the power consumption, to support power profiling of large scale HPC system. This measured data are communicated to a base station via Ethernet. This base station collects all such measurements which can be used for power profiling of IT load of the HPC system. This helps to provide better insight into the power utilization pattern. Wireless Sensor Network (WSN) is used to collect exhaust and inlet air temperature of the server nodes and this information is used for directing the cold air effectively. A vent control system is designed and fabricated for intelligently directing the air flow to the server node inlet. It takes the node power and the temperature data as its inputs. This enables supplying/ redirecting more cold air towards under-cooled nodes without creating an extra load on the cooling system, thereby bringing in effective cooling at reduced power consumption. As an added advantage this could help in hot-spots mitigation.","PeriodicalId":268630,"journal":{"name":"2016 2nd International Conference on Green High Performance Computing (ICGHPC)","volume":"316 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2016-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133567026","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}

引用次数: 2

Secure and multi-tenant Hadoop cluster - an experience 安全和多租户Hadoop集群——一种体验

2016 2nd International Conference on Green High Performance Computing (ICGHPC) Pub Date : 2016-02-01 DOI: 10.1109/ICGHPC.2016.7508069

Paresh Wankhede, Nayanjyoti Paul

{"title":"Secure and multi-tenant Hadoop cluster - an experience","authors":"Paresh Wankhede, Nayanjyoti Paul","doi":"10.1109/ICGHPC.2016.7508069","DOIUrl":"https://doi.org/10.1109/ICGHPC.2016.7508069","url":null,"abstract":"Data Analytics and Data Discovery are the most important facets in today's Business Domain where customer centric business decisions are the key. With ever increasing rate of data captivation, curation, management and requirement of data analytics, Hadoop has accounted itself as a major player in providing the data analytics and data processing backbone for any Organization that deals with ever increasing nuances of data management and processing. With every organizational setup of Hadoop clusters, we find it an ever increasing challenge to setup, manage and operate multiple Hadoop clusters, for managing different projects or managing different Tenants (clients). This results in a higher client onboarding time on Hadoop, cost of project ownership and effort to setup and manage separate clusters for separate projects/clients/tenants. However with the current trend of data security, companies are apprehensive of building a single large cluster and onboarding multiple clients on same common Hadoop cluster. This paper demonstrates how to set up a multi-tenant cluster which is big in size, scalable enough and has short client onboarding time without any client having access/knowledge/information of any other clients. Security features are also implemented on this multi-tenant cluster for authentication and authorization, so that only right client members have access to their allocated Hadoop resources like RAM, CPU and disk size. This paper also demonstrates how to create a fully functional and operational multi-tenant cluster with security at its core, reduced Cluster Management, higher data & resource security to provide an optimized Hadoop based solution offering in terms of cost and effectiveness.","PeriodicalId":268630,"journal":{"name":"2016 2nd International Conference on Green High Performance Computing (ICGHPC)","volume":"63 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2016-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131317070","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}

引用次数: 5

Efficient Key Management Protocol with Predictive Rekeying for Dynamic Networks 动态网络预测密钥更新的高效密钥管理协议

2016 2nd International Conference on Green High Performance Computing (ICGHPC) Pub Date : 2016-02-01 DOI: 10.1109/ICGHPC.2016.7508070

A. Prakash, V. R. Uthariaraj, L. Benjamin

{"title":"Efficient Key Management Protocol with Predictive Rekeying for Dynamic Networks","authors":"A. Prakash, V. R. Uthariaraj, L. Benjamin","doi":"10.1109/ICGHPC.2016.7508070","DOIUrl":"https://doi.org/10.1109/ICGHPC.2016.7508070","url":null,"abstract":"Transmission of datagrams to a group of hosts identified by a single destination address is Multicasting. Multicast is intended for group-oriented computing applications such as rescue teams, scientists with sharing of data and requirements for audio and video conferencing. All the multicast group members should have a group key to decrypt the datagram sent by the multicast server or source, which requires renewal for every membership view change to preserve forward and backward secrecy. Re-keying is the process in Key Management Protocol wherein a key distribution centre generates and distributes the cryptographic key during each membership event like member join or leave or transfer. The process of rekeying is required to preserve forward and backward secrecy, which induces computational, and communication overhead. In highly dynamic environments like wireless sensor networks, mobile clouds and IoTs, the members move from one area to another frequently and hence, this frequent join/leave of members in an area increases the rekey overhead. When the members become more dynamic, the rate of rekeying increases and hence the overhead increases. This paper proposes an Efficient Key Management Protocol with Predictive Rekeying Mechanism for Dynamic Networks, which reduces the effect of overhead per rekeying. The proposed mechanism reduces the frequency of rekeying by using a novel predictive rekeying, reduces the effect of mobility on the key management protocol, increases scalability, and reduces rekey overhead over other standard key establishment protocols.","PeriodicalId":268630,"journal":{"name":"2016 2nd International Conference on Green High Performance Computing (ICGHPC)","volume":"74 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2016-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122238691","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}

引用次数: 3

On the applicability of simple cache models for modern processors 简单缓存模型在现代处理器中的适用性

2016 2nd International Conference on Green High Performance Computing (ICGHPC) Pub Date : 2016-02-01 DOI: 10.1109/ICGHPC.2016.7508062

Rakhi Hemani, Subhasish Banerjee, Apala Guha

引用次数: 1