Comparison of fast, multiplatform XML libraries
Projects Main Page         Get PDF (Better formated)

1. The eXpat XML Parser (http://expat.sourceforge.net/)

2. The XML C Library for Gnome (http://www.xmlsoft.org/)

3. Apache Xerces for C++ parser (http://xml.apache.org/xerces-c/index.html)

4. Language Technology Group RXP XML parser (http://www.ltg.ed.ac.uk/software/xml/index.html)

5. Trolltech QT XML Module (http://doc.trolltech.com/3.1/xml.html)

6. IBM alphaWorks XML4C (http://www.alphaworks.ibm.com/tech/xml4c)

7. Oracle 9i XML Developers Kit for C/C++/Java (http://otn.oracle.com/tech/xml/xdk/)

8. Java XML pack (http://java.sun.com/xml/downloads/javaxmlpack.html)

1. The XSLT C Library for Gnome (http://xmlsoft.org/XSLT/) is distributed under MIT license and claims to support all systems, supported by XML C Library for Gnome.

2. Gnome DOM Engine (http://gdome2.cs.unibo.it/) is distributed under LGPL license and claims to support all systems, supported by XML C Library for Gnome.

3. XML Security Library (http://www.aleksey.com/xmlsec/) is distributed under MIT license and requires OpenSSL 0.9.6c or 0.9.7 (prefered).

4. Sablotron (http://www.gingerall.com/charlie/ga/xml/p_sab.xml) is distributed under GPL and Mozilla Open License 1.1 licenses. Sablotron is written in C++ and depends on expat XML parser.

5. Arabica (http://www.jezuk.co.uk/cgi-bin/view/arabica) is distributed under BSD license, written in C++ and claims to support Linux, Windows and different Unixes equiped with GNU compiler 3.

6. Xalan for C++ (http://xml.apache.org/xalan-c/index.html) is distributed under Apache license and known to work on Windows, Linux, Solaris, HP-UX and AIX platforms.

7. Pathan (http://software.decisionsoft.com/pathanIntro.html) is distributed under BSD style license, written in C++ and known to support Windows, Linux and Solaris platforms.

8. Center Point XML (http://www.cpointc.com/XML/) is distributed under Netscape Public License and claims to support many platforms, including Windows, Linux, BSD, Solaris, HP-UX, MacOS X. Written in C++.

9. IBM alphaWorks LotusXSL (http://www.alphaworks.ibm.com/tech/lotusxsl) is IBM shipped version of xalan library.

10. EasySOAP++ (http://easysoap.sourceforge.net/) is distributed under LGPL license, supports Linux and Windows platforms. And claims to be compatible with different SOAP implementations including, but not limited, Microsoft SOAP toolkit, Microsoft .NET platform, Apache, White Mesa, Phalanx.

11. iconv (http://www.gnu.org/software/libiconv/) is part of GNU c library and distributed under GPL license.

12. IBM International Components for Unicode (http://oss.software.ibm.com/icu/) is written in pure C unicde 3.2 library, provided under X Open Source license. ICU claims to support Linux, Windows and variety of Unix platforms.

13. Apache XML Security for C++ (http://xml.apache.org/security/index.html) is distributed under Apache license and known to work on Windows, Linux, Solaris, HP-UX and AIX platforms.

1. Expect DOM Modification part

Benchmark Setup

1. Hardware

1. Intel Pentium IV 2200 (22x100) MHZ (Nothwood)

2. Intel D850MV motherboard ( i850E + ICH2)

3. 4 x Samsung MR16R 256 MB PC800 Rambus DRAM

4. IBM GXP120 60GB hard drive

2. Software

1. DarkLin III (Mandrake 9.2 Based) Single User Mode

2. Kernel 2.4.22

3. GNU C Library 2.3.2

4. GNU C Compiler 3.3.1

5. Intel C++ Compiler 8.0

6. J2SE 1.4.2 + JAXP 1.2

3. Libraries

1. Expat 1.95.6

2. Sablotron 1.0.1

3. Arabica Jan04

4. CenterPoint XML 2.1.7

5. Gnome XML Library 2.6.5

6. Gnome XML DOM Library 0.8.1

7. Gnome XSLT Library 1.1.2

8. XML Security Library 1.2.4

9. Xerces for C++ 2.4.0

10. Xalan for C++ 1.7

11. LTG RXP Parser 1.2.8

12. Trolltech QT 3.3.0

13. Oracle XML Developers Kit for C 9.2.0.6.0

14. Oracle XML Developers Kit for Java 9.2.0.6.0

15. Apache XML Security for C++ 08.02.2004

16. Apache XML Security for Java 1.0.5D2

4. XML Files

1. XMLGEN - autogenerated simple XML file filled with random values, have 3 levels of deep and very simple XSD schema. Examples and schema can be accessed in XML Benchmark source package.

2. OPCGEN – autogenerated sequence of OPC(http://www.opcfoundation.org) messages filled with random data. This messages have 1-5 levels of deep and complex xsd schema. There are 16 different type messages which dramatically differs from each other. Examples and schema can be accessed in XML Benchmark source package.

3. 12 MB Resource Description File (RDF) from Dmoz.org

4. 1 MB Library Catalog file

Parsing Benchmarks

Validation Benchmark

XSLT Transformation Benchmark

DOM Benchmark

XML Security Benchmark

Stress Test

Parsers able process 64MB file: expat, arabica (SAX), cslxml (DOM,SAX), libxml (DOM,SAX), oracle (DOM,SAX), QT (SAX), RXP, Xerces (DOM,SAX), xml4c (DOM,SAX)

Parsers able process 128MB file: expat, arabica (SAX), cslxml (DOM,SAX), libxml (SAX), oracle (SAX), QT (SAX), Xerces (SAX), xml4c (SAX)

Parsers able process 256MB file: expat, arabica (SAX), cslxml (SAX), libxml (SAX), oracle (SAX), Xerces (SAX), xml4c (SAX)

Parsers able process 512MB file: Xerces (SAX), xml4c (SAX)

Standards Violations Notes

• Oracle bugs in schema verification and can drop core while validating non-valid documents.

• Gnome XML 2.6.5 library have problems with validation of xsd:QName attributes

• Xerces for C 2.4 don't catch schema implementation (xsi) type violations