wiki:sol4/HybridCloud/GEC22Report

Version 2 (modified by calyamp@missouri.edu, 5 years ago) (diff)

--

HybridCloud Project Status Report

Period: September 2014 - March 2015

I. Major accomplishments

A. Milestones achieved

  • Demonstrate two simulation-as-a-service apps co-operating with each other in a App Marketplace with any available GENI Rack
  • Report on usability studies with opt-in users in the form of case studies of user productivity, performance and collaboration experiences
  • Report on experiences implementing coupled app services on GENI, along with suggestions for improving GENI tools and services

B. Deliverables made

  • GEC22 Demo and Poster - Demonstrated multiple simulation-as-a-service apps co-operating with each other in a App Marketplace with an available GENI Rack
  • Talk at GEC22 Experimenter-Developer Roundtable - Title: User Opt-in Experiences in GENI Experiments
  • Technical Report on experiences implementing coupled app services on GENI - Title: App Chaining Software-as-a-Service for an Advanced Manufacturing Marketplace

II. Description of work performed during last quarter

A. Activities and findings

Activity: Address a salient limitation in App Marketplaces - the lack of a development environment that supports runtime capabilities for ‘Agile Manufacturing’ that efficiently integrates several Apps when building innovative products.
Findings: We investigated and developed a new Software-as-a-Service based App Runtime for the Marketplace environment that can be utilized for agile development of ‘Apps’ that involve high-performance modeling and simulation. Our solution approach features a web framework for the App runtime that: (a) builds upon the abstractions for most common workflows to support management of a generic set of ‘Apps’, (b) provides a chaining mechanism to create complex workflows for creation of new Apps based on customer requirements, and (c) runs complex simulation jobs on remote supercomputer resources and publishes customer-facing results for specific sets of inputs.

Activity: Leverage TotalSim's WheelSim product workflow and integrate it with GENI tools and services
Findings: We demonstrated how multiple Apps can be chained using our web framework for a product case study viz., ‘WheelSim’ deployed in the NSF GENI Cloud platform. Our results show improved App development convenience via rich UI elements interacting with RESTful web services and through dynamic chaining of workflows. Our study methodology also provides App developers with insights pertaining to estimation of resource cost, App pricing issues relevant to a manufacturing Marketplace, and the corresponding performance achievable with common App configurations in a cloud environment.

B. Project participants

The following are active participants in the project:

  • Prasad Calyam - PI
  • Ray Leto - Advanced Manufacturing Expert/Subcontractor
  • Amit Kumar Akula - Research Assistant
  • Ronny Bazan Antequera - Research Assistant

C. Publications (individual and organizational)

  • S. Ramisetty, P. Calyam, J. Cecil, A. Akula, R. Antequera, R. Leto, “Ontology Integration for Advanced Manufacturing Collaboration in Cloud Platforms”, IFIP/IEEE Intl. Symposium on Integrated Network Management (IM), 2015.
  • A. Akula, P. Calyam, R. Antequera, S. Ramisetty, R. Leto, “Advanced Manufacturing Collaboration in a Hybrid Cloud-based App Marketplace”, MU Technical Report, 2015.

D. Outreach activities

  • P. Calyam, “Simulation-as-a-Service: A GENI Experiment for Advanced Manufacturing”, Invited Talk at Next-gen Applications - 1 Billion Bits, Kansas City, KC, Fall 2014.
  • P. Calyam, “Cloud Computing as a City-supported Utility”, MU Computer Science Seminar Series, Fall 2014.
  • MU Cloud Computing I Class Project, Spring 2015; Title - "Hosting and Pricing of an Advanced Manufacturing App Marketplace"
  • P. Calyam, "Using GENI in Large Classes within a Cloud Computing course", GEC22 Education Session Talk, 2015.
  • GEC22 Demo Video

E. Collaborations

Discussion of our contributions with collaborators at Ohio Supercomputer Center and Oklahoma State University.

F. Other Contributions

None.