We have identifi\fed three major work directions to be investigated as part of DALHIS associated team for 2015:<\/p>\n
Scienti\ffic Workows<\/strong><\/span> LBNL<\/strong><\/p>\n Joint work<\/strong><\/p>\n Energy-effi\u000ecient cloud elasticity for data-driven applications<\/strong><\/span> We have identifi\fed three major work directions to be investigated as part of DALHIS associated team for 2015: Scienti\ffic Workows The development of the integrated workfow engine using HOCL and Tigres continues. We recently started to explore the expressiveness of HOCL to express self-adaptive behaviors.\u00a0 Tigres will speci\ffically focus next year on failure recovery and … <\/p>\n
\nThe development of the integrated workfow engine using HOCL and Tigres continues. We recently started to explore the expressiveness of HOCL to express self-adaptive behaviors.\u00a0 Tigres will speci\ffically focus next year on failure recovery and fault tolerance API. Additionally, Tigres will investigate decentralized execution
\nand optimizations to enable executions on next-generation HPC systems that have deeper I\/O-memory hierarchies. Our work plan is organized as follows.
\nMyriads<\/strong><\/p>\n\n
\n
\n
\n
\n
\n
\n
\n
\n
\n
\nDistributed and parallel systems o\u000ber to users tremendous computing capacities. They rely on distributed
\ncomputing resources linked by networks. They require algorithms and protocols to manage these resources
\nin a transparent way for users. Recently, the maturity of virtualization techniques has allowed for
\nthe emergence of virtualized infrastructures (Clouds). These infrastructures provide resources to users
\ndynamically, and adapted to their needs. By benefi\fting from economies of scale, Clouds can effi\u000eciently
\nmanage and off\u000ber virtually unlimited numbers of resources, reducing the costs for users.
\nHowever, the rapid growth for Cloud demands leads to a preoccupying and uncontrolled increase of
\ntheir electric consumption. In this context, we will focus on data driven applications which require to
\nprocess large amounts of data. These applications have elastic needs in terms of computing resources
\nas their workload varies over time. While reducing energy consumption and improving performance are
\northogonal goals, this internship aims at studying possible trade-o\u000bffs for energy-effi\u000ecient data processing
\nwithout performance impact. As elasticity comes at a cost of recon\fgurations, these trade-off\u000bs will
\nconsider the time and energy required by the infrastructure to dynamically adapt the resources to the
\napplication needs.
\nThe validations of the proposed algorithms may rely on the French experimental platform named
\nGrid’5000. This platform comprises about 8,000 cores geographically distributed in 10 sites linked with
\na dedicated gigabit network. Some of these sites have wattmeters which provide the consumption of the
\ncomputing nodes in real-time. This validation step is essential as it will ensure that the selected criteria
\nare well observed: energy-effi\u000eciency, performance and elasticity.
\nData Ecosystem<\/span><\/strong>
\nScienti\fc now routinely generates large and complex datasets as the result of experiments, observations,
\nor simulations and the number of scientists analyzing these datasets are also growing. These datasets
\nand their relationships have become increasingly di\u000efficult to manage and analyze. We will continue our
\nwork in identifying, developing and implementing the data ecosystem for scienti\ffic applications.
\nIn addition to the above activities, we will continue our work on FRIEDA to explore elasticity of
\ndata and its impact on execution. Elasticity and auto-scaling of compute resources has been explored
\nbefore. However, its interaction with storage and data management is not well understood. There are a
\nnumber of open issues in this space. First, it is unclear how growing or shrinking data volumes should
\nbe managed in virtualized environments. Also, when a compute resource is brought up or removed from
\nthe pool, how should the data on these volumes be managed. In the context of FRIEDA, we will design,
\nimplement and evaluate data management strategies to manage elasticity. This work will be evaluated
\non Amazon, Grid 5000 and other cloud testbeds.
\nThe DALHIS associated team has been in deep discussions about the appropriate data ecosystem
\nand its components suitable for scienti\fc applications. In the coming year, we will be writing the paper
\noutlining what a data ecosystem accounts for.<\/p>\n","protected":false},"excerpt":{"rendered":"