EGIIn SPIRE Virtual Research Environment Optimal Engineering Design

  • Slides: 36
Download presentation
EGI-In. SPIRE Virtual Research Environment “Optimal Engineering Design as a Service” Prof. Anatolii Petrenko

EGI-In. SPIRE Virtual Research Environment “Optimal Engineering Design as a Service” Prof. Anatolii Petrenko petrenko@cad. kiev. ua , www. cad. kpi. ua Tel. +38 -044 -280 -9046, fax: +38 -044 -280 -8771 System Design (CAD) Dept. , National Technical University of Ukraine ”KPI” EGI-In. SPIRE RI-261323 www. egi. eu

Agenda Needed User Community “Engineering” Service Oriented Architecture (SOA) Cloud Computing and SOA Service

Agenda Needed User Community “Engineering” Service Oriented Architecture (SOA) Cloud Computing and SOA Service Oriented Computing (SOC) SOC Applications SOC in Engineering Design Conclusions EGI-In. SPIRE RI-261323 2 www. egi. eu

SOA Definition • By definition, a service is performance of work by one for

SOA Definition • By definition, a service is performance of work by one for another. Services are discrete pieces of software, implemented using well-defined standards and interface specifications. A service, once developed and tested is submitted to a directory or registry so that it is available to other developers. • Service Oriented Architecture (SOA) provides a looselyintegrated suite of services that can be used within multiple business domains. It is a flexible set of design principles used during the phases of systems development and integration. 3 EGI-In. SPIRE RI-261323 www. egi. eu

How services encapsulate logic 4 EGI-In. SPIRE RI-261323 www. egi. eu

How services encapsulate logic 4 EGI-In. SPIRE RI-261323 www. egi. eu

Cloud computing and SOA ³ Cloud Computing (CC): an approach to a shared IT

Cloud computing and SOA ³ Cloud Computing (CC): an approach to a shared IT infrastructure in which large pools of computer systems are linked together to provide IT services on-demand, like a public utility. • Cloud Computing Services are classified as: Cloud Infrastructure, Cloud Storage, Cloud Platform, Cloud Applications. • SOA implementations are primarily for exchange of information between systems and a network of systems. On the other hand, Cloud computing focuses on leveraging the network for the entire gamut of IT functions. 5 EGI-In. SPIRE RI-261323 www. egi. eu

Combining SOA and CC Service-oriented computing (SOC) represents a new generation distributed computing platform.

Combining SOA and CC Service-oriented computing (SOC) represents a new generation distributed computing platform. § Service-oriented computing and Cloud computing have a reciprocal relationship — one provides the computing of services, and the other provides the services of computing. § Developers can create new applications quickly by integrating available services which are independent entities and can be invoked at run-time without any knowledge of the platform or the programming language. Modern interdisciplinary CAE/CAD software capable to utilize remote computing resources to solve engineering problems of high complexity and with desired accuracy at the reasonable simulation time are of high demand today. 6 EGI-In. SPIRE RI-261323 www. egi. eu

SOC and Cloud Computing Neal Leavitt, Yankee Group, IEEE Computer Society, January 2009 EGI-In.

SOC and Cloud Computing Neal Leavitt, Yankee Group, IEEE Computer Society, January 2009 EGI-In. SPIRE RI-261323 7 www. egi. eu

User EGI Communities EGI-In. SPIRE RI-261323 8 www. egi. eu

User EGI Communities EGI-In. SPIRE RI-261323 8 www. egi. eu

Main objectives of the project • To satisfy the requirements of grid users in

Main objectives of the project • To satisfy the requirements of grid users in scientific and applied research tools which support networked collective design and provide analysis of complex objects using grid/cloud facilities. • To be functionality distributed across the ecosystem of both web services and grid/cloud services. • To provide a workflow editor for composition and orchestration of these services into a user defined computing scenario or a design route. • To be based on innovative numeric methods of parallel of mathematical modeling and optimal design. EGI-In. SPIRE RI-261323 9 www. egi. eu

Engineering Platform Architecture 10 EGI-In. SPIRE RI-261323 www. egi. eu

Engineering Platform Architecture 10 EGI-In. SPIRE RI-261323 www. egi. eu

Platform Supporting Services • Services of management and hosting (e. g. cloud hosting, event

Platform Supporting Services • Services of management and hosting (e. g. cloud hosting, event processing and management, mediation and data services, security, messaging, storage , visualization, network connection, etc. ). • A composition editor and execution engine that allows end users with little or no programming skills to create and run a composite web application front-end as a mashup built from widgets and operators • Location management of mobile devices via A-GPS, CID and Wi. Fi • Composition Engine is a toolset for creating and executing composed services and applications • A Registry acts as a universal directory of information used for the maintenance, administration, deployment and retrieval of services 11 EGI-In. SPIRE RI-261323 www. egi. eu

Platform Supporting Services 12 EGI-In. SPIRE RI-261323 www. egi. eu

Platform Supporting Services 12 EGI-In. SPIRE RI-261323 www. egi. eu

Application support services • to investigate the generalized engineering design process and to select

Application support services • to investigate the generalized engineering design process and to select its loosely coupled stages and procedures for subsequent their transferring to the forms of standardized web-services; • to analyze the existing mathematical modeling and optimal design software for the possible re-use of the best algorithms and design procedures implementations in the creating the depository of applied web services; • to develop a container with interfaces for standardized individual web-services based on international standards and protocols which allow building compositions from these web-services as design (calculations) workflows. • to implement novel service-oriented design paradigm in developing Engineering Design Platform, in particular, for modeling and optimization of Nonlinear Dynamic Systems, based on components of different physical nature and being widely spread in different scientific and engineering fields. 13 EGI-In. SPIRE RI-261323 www. egi. eu

Engineering in Cloud 14 EGI-In. SPIRE RI-261323 www. egi. eu

Engineering in Cloud 14 EGI-In. SPIRE RI-261323 www. egi. eu

An Engineer in the Platform EGI-In. SPIRE RI-261323 15 www. egi. eu

An Engineer in the Platform EGI-In. SPIRE RI-261323 15 www. egi. eu

Web. ALLTED – All Technology Designer • A prototype is native Web. ALLTED (ALL

Web. ALLTED – All Technology Designer • A prototype is native Web. ALLTED (ALL TEchnology Designer) software. It was developed not only for simulation and analysis, but for providing project procedures such as: • • • parametric optimization tasks; optimal tolerance assignments; centering availability regions; yield maximization; design of Nonlinear Dynamic Systems composed of either/and electronic, hydraulic, pneumatic, mechanical, electromagnetic, and other elements. • Web. ALLTED is especially useful in the development of new products which combine various physical phenomena in one device or mechanical, electrical, thermal, optical and other components EGI-In. SPIRE RI-261323 16 www. egi. eu

Web. ALLTED usage examples EGI-In. SPIRE RI-261323 17 www. egi. eu

Web. ALLTED usage examples EGI-In. SPIRE RI-261323 17 www. egi. eu

Web. ALLTED usage examples EGI-In. SPIRE RI-261323 18 www. egi. eu

Web. ALLTED usage examples EGI-In. SPIRE RI-261323 18 www. egi. eu

Web. ALLTED usage examples EGI-In. SPIRE RI-261323 19 www. egi. eu

Web. ALLTED usage examples EGI-In. SPIRE RI-261323 19 www. egi. eu

Web. ALLTED usage examples EGI-In. SPIRE RI-261323 20 www. egi. eu

Web. ALLTED usage examples EGI-In. SPIRE RI-261323 20 www. egi. eu

Web. ALLTED usage examples EGI-In. SPIRE RI-261323 21 www. egi. eu

Web. ALLTED usage examples EGI-In. SPIRE RI-261323 21 www. egi. eu

Web. ALLTED usage examples EGI-In. SPIRE RI-261323 22 www. egi. eu

Web. ALLTED usage examples EGI-In. SPIRE RI-261323 22 www. egi. eu

System on a Chip EGI-In. SPIRE RI-261323 23 www. egi. eu

System on a Chip EGI-In. SPIRE RI-261323 23 www. egi. eu

Web. ALLTED exclusively offers: • Automatic object's mathematical models generation; • Reducing the dimension

Web. ALLTED exclusively offers: • Automatic object's mathematical models generation; • Reducing the dimension of the formed mathematical model; • Faster simulation speed and improved numerical convergence for DC, AC and TR Analyses; • Sensitivity analysis for frequency and transient analyses; • Comprehensive optimization procedure and optimal tolerances assignment ; • Alternative approach to the secondary response parameters determination (delays, rise and fall times, etc. ); • Powerful user-defined modeling capability. • Original way of generating a system-level model of MEMS from FEM component equations. EGI-In. SPIRE RI-261323 24 www. egi. eu

Novel numerical methods • The new method for Steady State (DC) analysis which provides

Novel numerical methods • The new method for Steady State (DC) analysis which provides the quick descent to the solution point region from any starting point. • The Diagonal Modification Method which helps considerably preserve convergence of linearized equations solution without re-ordering when matrix element values change from one iteration to another iteration. • The Optimization Variable-order Methods which is equivalent to taking into consideration five terms of Tailor’s series for the Goal function which considerably improve determination of a direction to the optimal point. • The Optimal Tolerances Assignment Method which is based on applying Optimization procedures and takes into account the prescribed deviations of Controlled Output Parameters • The Implicit Linear Multi-step Variable-order Integration Method for Transient Analysis(TR) which uses high order back differences that allows to select the proper one resulting in minimization of solution time for prescribed accuracy. EGI-In. SPIRE RI-261323 25 www. egi. eu

Novel numerical methods . • Statistical Yield Maximization Method which provides “centering” the solution

Novel numerical methods . • Statistical Yield Maximization Method which provides “centering” the solution point in the region of acceptable solutions • References: ▪ Petrenko A. , Sigorsky V. , “Algorithmic analysis of electronic circuits", Western Periodical Corp. , San Francisco, 1975, 618 p. ▪ Petrenko A. , Ladogubets V. , Tchkalov V. , Pudlowski Z. ”ALLTED-a Computer-Aided System for Electronic Circuit Design”, UICEE (UNESCO), Melbourne, 1997, 204 p. ▪ Michail Zgurovsky, Anatoly Petrenko, Volodymyr Ladogubets, Oleksii Finogenov, Bogdan Bulakh. “Web. ALLTED: Interdisciplinary Simulation in Grid and Cloud” , Computer Science (Cracow), 14 (2) 2013, pp. 295306. EGI-In. SPIRE RI-261323 26 www. egi. eu

Simulation Workflow • Workflow concept • Design scenario can be represented as a sequence

Simulation Workflow • Workflow concept • Design scenario can be represented as a sequence of analysis procedures and data management activities (preparation, transformation, visualization, backup, transferring, encoding etc. ) • Thus, design scenarios can be treated as design workflows: Math. model composition EGI-In. SPIRE RI-261323 Static analysis Transient analysis Optimization procedure A 27 www. egi. eu

Application services • Each computing step in a workflow invokes specific webservice (grid/cloud-service) •

Application services • Each computing step in a workflow invokes specific webservice (grid/cloud-service) • Web services are representing the basic building blocks of system's functionality: input data preprocessing; mathematical model development and its dimension reduction; DC, AC, TR and sensitivities analysis; parametrical optimization, tolerances assignment; statistical analysis and yield maximization; results processing etc. • Service orchestration based on WS-BPEL -2 • Semantic Web technologies also can be leveraged (automatic service discovery, composition etc. ) EGI-In. SPIRE RI-261323 28 www. egi. eu

General Architecture WWW PROJECTS PORTAL SERVER GRAPH. EDITOR TASK SUBMISSION & MONITORING WORKFLOWS EXEC.

General Architecture WWW PROJECTS PORTAL SERVER GRAPH. EDITOR TASK SUBMISSION & MONITORING WORKFLOWS EXEC. SERVER WEB SERVER RESOURCES EGI-In. SPIRE RI-261323 WS-ORCHESTRATION ENGINE WEB SERVICE HP HP WF. MANAGER GRID SERVICE G WEB SERVER G 29 www. egi. eu

Data diagram XML file 1. User creates workflow graph from registered components in graph

Data diagram XML file 1. User creates workflow graph from registered components in graph editor 2. Abstract workflow XMLbased description (specifying workflow items, their dependencies, input and output ports, input data) is generated and then submitted by task submission module Workflow description generator BPEL-process generator Service registry EGI-In. SPIRE RI-261323 3. Concrete BPEL-process description is generated by workflow management service for further execution by orchestration engine BPELengine 30 www. egi. eu

Testing Details Testbed is residing on KPI HPC Center computing resources § 32 cores

Testing Details Testbed is residing on KPI HPC Center computing resources § 32 cores (4 nodes with 2 x Intel Xeon E 5345 2, 33 GHz 4 Cores, 8 Gb RAM) § OS Centos, LRMS Torque 2. 3. 6 + Maui 3. 3 scheduler § Grid middleware Nordugrid ARC 0. 8. 3 EGI-In. SPIRE RI-261323 Example of “stiff” circuit r 1=r 2=r 3=1 Ω, с1=0. 1 F, с2=10 F, с3=1 Е-5 F 31 www. egi. eu

Web. ALLTED 32 EGI-In. SPIRE RI-261323 www. egi. eu

Web. ALLTED 32 EGI-In. SPIRE RI-261323 www. egi. eu

Schematic and Workflow Editors EGI-In. SPIRE RI-261323 33 www. egi. eu

Schematic and Workflow Editors EGI-In. SPIRE RI-261323 33 www. egi. eu

Screenshots of simulation EGI-In. SPIRE RI-261323 34 www. egi. eu

Screenshots of simulation EGI-In. SPIRE RI-261323 34 www. egi. eu

Conclusions • The multi-layered architecture of the cloud-enabled computer simulation software was presented. This

Conclusions • The multi-layered architecture of the cloud-enabled computer simulation software was presented. This architecture is characterized by the following: • functionality is distributed across the ecosystem of both web services and grid/cloud services (enabling utilization of grid/cloud computing resources) • it is compatible with adopted standards and protocols • it supports custom user analysis scenario development and execution • functionality is accessible with lightweight web interface • it hides the complexity of web-service interaction from user with abstract workflow concept and simple graphical workflow editor. • We are open for cooperation especially for moving this software to Cloud within EU Projects like EGI – In. SPIRE, VENUS-C, Stratus. Lab, m. OSAIC and HELIX NEBULA (Science Cloud) EGI-In. SPIRE RI-261323 35 www. egi. eu

Any questions? Thanks you! 36 EGI-In. SPIRE RI-261323 www. egi. eu

Any questions? Thanks you! 36 EGI-In. SPIRE RI-261323 www. egi. eu