Below is a list of public deliverables that were produced throughout the duration of the CloudLightning Project. At each deadline or milestone, the deliverables were made available for download and public consumption.
WP2 – Use Cases and Testbed
D2.1.1. USE CASE REQUIREMENTS
This report presents four use cases that are targeted in the CloudLightning project: genomics, oil and gas exploration, ray tracing, and self-optimised libraries. The report presents a general overview over the HPC market, details the requirements, stakeholders and means to evaluate the identified use case applications. It also outlines further application areas in sparse and dense matrix computation.
D2.2.1. INTEGRATED USE CASES
This deliverable captures the steps followed in converting the service specifications developed in D2.1.1 into completely automated deployable services that can be readily deployed in the CL heterogeneous resource pool.
WP3 – Architecture, Integration and Evaluation
D3.1.2. Report on State of the Art and Draft Architecture
This deliverable reports on state of the art and draft architecture, specifying the detailed architecture of the system’s software components.
WP4 – Self-organisation and Self-management
D4.1.1. Protocol Specification and API
UCC, in collaboration with the University of Central Florida, have published simulation results that demonstrate that core-periphery arrangement of independent computing resources can self-organise in a scalable fashion to meet service requests. This report produces a SoA-based network protocol and implementation for self-organisation, self-management of cloud infrastructures. Existing technologies, such as WS-Discovery and ReSTful web services are evaluated and leveraged where appropriate. The protocol and its implementation are modular, allowing for the basic protocol to be extended to support a variety of heterogeneous co-processor types.
D4.2.1. Local Decision Strategies
This report devises and evaluates algorithms that make local decisions on cloud resources with the goal of optimising global cloud metrics such as energy usage. These algorithms encode decisions such as whether or not the node should join a coalition that is being advertised and whether the node should attempt to offload its workload and enter a sleep state. Simulation is used as appropriate to evaluate the strategies. The collection of chosen algorithms is used to implement the set of default behaviours for the self-organisation and self-management protocol.
D4.4.1. Service Deployment and Management
This deliverable presents an integrated telemetry and monitoring system design and its implementation procedures for the CloudLightning project. The proposed system is able to collect, process, store and distribute the CL resource metrics among other systems which are requiring telemetry data in a flexible and scalable manner.
WP5 – Service Description Language and Gateway Service
D5.1.1. Service Description Format
The focus of this deliverable is to perform state of the art analysis of existing service description languages (TOSCA, USDL) as well as standardisation activities. Additional effort has been put into surveying specifications coming from related FP7 projects such as MODAClouds. Requirements for the service description language are drawn in conjunction with the use cases that the project will service.
D5.2.1. Gateway Service
In the context of creating a self-organizing and self-managing cloud infrastructure that is easily accessible by non-technical users we propose to construct a Service Gateway that enables ease of communication between the two entities. Gateway Service is intended to facilitate the access to heterogeneous resources exposed by the CloudLightning system and the world.
WP7 – Large Scale Simulation
D7.1.1. Large scale modelling and simulation framework
This deliverable focuses on the implementation of the CloudLightning simulation framework,
a generalized and extensible simulation framework that enables the seamless simulation
and experimentation of emerging Cloud computing infrastructures and HPC applications.
D7.2.1. Simulation execution and evaluation
This deliverable focuses on implementation details for the CloudLightning simulation
platform and its parameterized execution on distributed memory systems. The source
code of the platform is thoroughly analyzed and the interaction between classes and
objects through the simulation process is discussed. The execution of the platform on a
distributed system is also performed, using a variety of simulation inputs. The numerical
results show high scalability of the platform in terms of the number of cloud servers and
tasks simulated, both when the traditional cloud delivery system or the Self-Organization
and Self-Management resource allocation scheme were used. An evaluation roadmap is
also discussed, focusing on the main criteria that will be used for the evaluation of the
CloudLightning approach through the simulation platform.
D7.3.1. Simulation report
This deliverable focuses on experimentation with the CloudLightning simulation framework
developed under Task 7.2. The simulation executions were parameterized according
to the use case executions on the CloudLightning testbed, while the simulation models
were verified using traces from the testbed. The simulation framework was also verified
against other open source simulation frameworks, for validation of the accuracy of the
simulation results. The exhaustive experimentation performed, includes executions with
different numbers of incoming cloud tasks and size of the cloud, both using traditional
centralized and Self-Management and Self-Organization resource allocation mechanisms.
WP8 – Exploitation and Dissemination
D8.1.1. Exploitation Plan and Strategic Guidelines for Exploitation
The first deliverable provides details about the exploitation plan and Strategic Guidelines for exploitation. Follow-up deliverables will provide updates on targets, detailed figures and progress towards implementing the exploitation strategy. A separate report will be issued in M30 relating to the outputs from the Delphi Method research (D8.1.4). The final report will include a review of all exploitation activities undertaken during the project.
D8.1.2. Exploitation Progress Report
This deliverable reports on actions taken and impact achieved with regard to exploitation activities in Y2, defines individual exploitation strategy for the CloudLightning consortium partners and outlines the exploitation activities expected during Y3.
D8.2.1. Dissemination Plan
The first deliverable provides details about the Dissemination Plan, implementation of the project website, Basecamp and social networking accounts and a progress report on agreed activities to raise public awareness and participation, and impact against planned metrics in the first 12 months. Follow-up deliverables will provide updates on targets, detailed figures and progress towards implementing the dissemination strategy. The final report will include a review of all dissemination activities undertaken during the project.
D8.2.2. Dissemination Progress Report
This document updates the initial Dissemination Plan (D8.2.1), reports on actions taken and impact achieved with regard to dissemination and communication activities in Y2, and outlines dissemination strategy for Y3.
D8.3.1. Concertation Plan
The deliverable provides details about the Concertation Plan and a progress report on concertation activities and impact against planned metrics in the first 12 months. Follow-up deliverables will provide updates on targets, detailed figures and progress towards implementing the concertation plan. The final report will include a review of all concertation activities undertaken during the project.
D8.3.2. Concertation Progress Report
The purpose of this document is to report on actions taken and impact achieved with regard to concertation and orchestration activities in Y2, update the initial concertation plan (D8.3.1) produced in January 2015 and outline concertation strategy for Y3.