Cloud versus Cloud How Will Cloud Computing Shape

  • Slides: 3
Download presentation
Cloud versus Cloud: How Will Cloud Computing Shape Our World? Panel HPDC Chicago Illinois

Cloud versus Cloud: How Will Cloud Computing Shape Our World? Panel HPDC Chicago Illinois June 24 2010 http: //hpdc 2010. eecs. northwestern. edu June 24 2010 Geoffrey Fox [email protected] edu http: //www. infomall. org http: //www. futuregrid. org Director, Digital Science Center, Pervasive Technology Institute Associate Dean for Research and Graduate Studies, School of Informatics and Computing Indiana University Bloomington

When and How are Clouds Useful • Expect Clouds to excel in “high throughput

When and How are Clouds Useful • Expect Clouds to excel in “high throughput computing” but not replace high end Compute Grids – certainly won’t replace exascale or petascale systems – Clouds are (by definition) commercially supported approach to large scale computing while current Grid technology involves “non -commercial” software solutions which are hard to evolve/sustain • Services still are correct architecture with either REST (Web 2. 0) or Web Services – we should always offer X as a service where ever possible as in SQL as a Service • Workflow will move smoothly from Grids to Clouds • Platform as a Service should motivate new generation of Scientific Computing Platforms

Features of a Scientific Computing Platform? Blob: Basic storage concept similar to Azure Blob

Features of a Scientific Computing Platform? Blob: Basic storage concept similar to Azure Blob or Amazon S 3; Disks as a service as opposed to disks as a Drive/Elastic Block Store DPFS Data Parallel File System: Support of file systems like Google (Map. Reduce), HDFS (Hadoop) or Cosmos (dryad) with compute-data affinity optimized for data processing Table: Support of Table Data structures modeled on Apache Hbase or Amazon Simple. DB/Azure Table. Bigtable (Hbase) v Littletable (Document store such as Couch. DB) Roles: Web Role automates portals? Worker role dynamic computing Queues: Publish Subscribe based queuing system Map. Reduce: Support Map. Reduce Programming model including Hadoop on Linux/Amazon, Dryad on Windows HPCS and Twister on Windows, Linux, Azure, Amazon