Research Seminar Robert van Engelen Computer Science & CSIT Florida State University

Research, Nov 4,  Restructuring compiler analysis and transformation  Compilation techniques for embedded systems  Web and Grid services technologies  Bioinformatics and life sciences  Atmospheric and oceanographic sciences Research Projects

Research, Nov 4, Current Funding Support Improving Symbolic Analysis of Restructuring Compilers DOE Early Career PI Program (Compilers & Algorithms for High-Performance Computing, Networking, and Data Management) Automatic Validation of Code Improving Transformations and Related Applications (with Dr. Whalley) Collaborative Research in Compilers (with Dr. Whalley)

Research, Nov 4, Research Assistants  Chris Baker (for MS) collab. with Dr. Gallivan  Compiler technology & numerical algorithms  Johnnie Birch (for MS & PhD)  Compiler analysis techniques  Burt Walsh (for PhD)  Compiler & embedded systems technology  Yixin Shou (for PhD)  Compiler technology  Open position (for PhD)  Web and grid services technologies

Research, Nov 4, Web and Grid Services  Web and Grid services are remote applications, e.g. on the Web, that provide various data, document, and compute resources to client applications  Web/Grid service technology is a big step forwards in distributed computing on the Internet  Bridges platform and language interop gaps

Research, Nov 4, Web/Grid Service Protocols  Web/Grid services technologies utilize the SOAP and WSDL protocols  SOAP (Simple Object Access Protocol) is a light-weight protocol based on XML as the marshalling format for request and response messages, typically with HTTP  WSDL is an XML-based Web/Grid services description language

Research, Nov 4, WSDL 1. A Web Service registers its existence in a registry 2. The Web Service publishes its WSDL in the registry 3. SOAP remote procedure calling by clients developed from this WSDL

Research, Nov 4, gSOAP Web Services  SOAP/XML language binding for C & C++  SOAP/XML capabilities build into the C and C++ language  Uses compiler technology  Based on RPC compiler  No need to write any wrapper code  Generates SOAP/XML parsing routines (no DOM or SAX parser required)

Research, Nov 4, gSOAP Web Services  gSOAP is open source (SourceForge project) (C/C++, Java)  Platform independent (now with automake/conf)  Integrated TCP/IP, HTTP, DIME, SOAP, and XML stacks  Designed for high-performance throughput

Research, Nov 4, gSOAP Features  SOAP 1.1 and 1.2 compliant  WSDL 1.1 import and export  Stand-alone servers (HTTP 1.0 and partially 1.1)  Supports keep-alive and chunked transfers  CGI, Fast-CGI, Apache_mod, IIS  DIME attachments  Supports pure C or C/C++ application mix  Multi-threaded Web services, MT safe  Extensible  Plug-ins, callbacks, and user-defined (de)serializers

Research, Nov 4, gSOAP Users in Industry  Adobe Systems (eg. GoLive Web site development software)  IBM Alphaworks  Xerox  CoreNetworks  Powerllel  Siemens  Ericsson  Globus toolkit (DOE, Argonne National Labs)  Harness toolkit (DOE)  …

Research, Nov 4, Native C & C++ Application Code to Serve Requests (RPC Parameter Marshalling is Fully Automatic) libgsoap.a gSOAP Runtime Environment (Memory & I/O Management) Web Service Application Implementing a Web Service Application myService.h Specification of Remote Procedures and Parameter Types gSOAP RPC Compiler myService.wsdl soapService.c soapC.c RPC Skeleton Routines and XML Parsers/Generators Publish WSDL

Research, Nov 4, Native C & C++ Application Codes with SOAP RPC Calls (RPC Calls and Parameter Marshalling are Automatic) Client Application libgsoap.a gSOAP Runtime Environment (Memory & I/O Management) Implementing a Client Application someService.wsdl gSOAP WSDL importer WSDL Publication soapClient.c soapC.c RPC Stub Routines and XML Parsers/Generators gSOAP RPC Compiler someService.h Specification of Remote Procedures and Parameter Types

Research, Nov 4, SOAP RPC Marshalling struct BG { int val; struct BG *left; struct BG *right; }; int remoteMethod(struct BG input, struct BG *output);

Research, Nov 4, SOAP RPC Demarshalling

Research, Nov 4, Application to Life Sciences  Determine the “Tree of Life” from sequenced genome data  Serialize and deserialize phylogenetic trees in XML  Massive parallel computations  Cluster of workstations SOAP/XML

Research, Nov 4, Results 1)Interoperability Testing 2)Legacy C Code Web Service Integration 3)Scalability and Performance

Research, Nov 4, Interoperability Testing  WhiteMesa.org “interop lab”  Suite of test cases designed for real-time interoperability testing  Interop with: Apache Axis,.NET, Delphi,…

Research, Nov 4, Legacy C Code Integration: Linear System Solver Service LU decomposition from “Numerical Recipes in C”: ludcmp(double **a, int n, int *indx, double *d); ludcmp(double **a, int n, int *indx, double *d); Service routine (array size n stored in structs for vectors and matrices) ludcmp /**** ludcmp request ****/ ( struct mat *a, /* input: matrix */ struct ludcmpResponse /**** ludcmp response ****/ { struct mat *a; /* output: matrix (decomposed) */ struct ivec *indx; /* output: reordering vector */ double *d; /* output: arg for determinant */ } *result ); Struct declarations: struct vec /**** vector ****/ { double *__ptr; /* pointer to array of double */ int __size; }; /* run-time array size */ struct mat /**** matrix ****/ { struct vec *__ptr; /* pointer to array of vectors */ int __size; }; /* run-time array size */

Research, Nov 4, Linear System Solver Service Linear system solver service (LSSS) code (CGI-based): main() { soap_serve(soap_new()); /* process request */ } Example LSSS client application code: main() { struct soap soap; /* runtime environment */ struct mat a; /* input matrix */ struct ludcmpResponse result; /* result output */ … soap_init(&soap); /* init runtime */ if (soap_call_ns__ludcmp(&soap, “URL”, “”, &a, &result)) soap_print_fault(&soap, stderr); … … = result.a.__ptr[i]->__ptr[j]; /* index a[i,j] */ }

Research, Nov 4, Scalability and Performance 1)Scalability and overhead of communication vs. computation  LU-based double fp. matrix inversion 2)Performance (send 32bit int matrix)  Full SOAP XML-encoded int32 matrix  SOAP Base64-encoded int32 matrix  CGI-based SOAP Web Service  Stand-alone SOAP Web Service  Java RMI

Research, Nov 4, Communication Overhead: Matrix Inversion  Stand-alone linear system solver service  Full double fp. Matrix representation in SOAP XML  Total time of client request and server response (100BaseT, Dual PIII 550MHz, Red Hat Linux)

Research, Nov 4, Full XML Versus Base64 int32 Matrix Representations  Total time of client request and server response (100BaseT, Dual PIII 550MHz, Red Hat Linux)  SOAP XML- encoded 32bit int matrix  SOAP Base64- encoded 32bit int matrix

Research, Nov 4, XML vs. Base64 vs. Java RMI with Stand-Alone Service  Total time of client request and server response (100BaseT, Dual PIII 550MHz, Red Hat Linux)  SOAP XML- encoded 32bit int matrix  SOAP Base64- encoded 32bit int matrix  Java RMI (1.2.2)

Research, Nov 4, Questions?