GCN Tech Blog

By GCN Staff

Blog archive

Revisiting XML compression

In the most recent issue, we reported on how the Efficient XML Interchange (EXI) framework could cut back on the size of Extensible Markup Language-based event streams.

We stated that it was the only XML-compression standard, which, as it turns out, is wrong.

It was OSS Nokalva's Alessandro Triglia who corrected us on this matter. OSS Nokalva has a set of Fast Infoset tools.

Since Triglia who also had a lot interesting to say about Fast Infoset, and even made a few revealing comparisons with EXI, we'll reprint his e-mail here:


In your article you state that 'Many companies offer network appliances ...but no open standard has addressed this issue.' This is incorrect. Fast Infoset is an ISO standard (ISO/IEC 24824-1) which does precisely that, and was published in 2005 by both ISO/IEC and ITU. The standard is publicly available and can be downloaded for free from the ITU website.

It is a completely open standard which doesn't come with any royalties or patents or other encumbrances. There are open-source implementations as well as commercial implementations available. Implementation is relatively easy, as one of the design goals of this standard was ease of implementation.

Let me also point out that the statement in your article about EXI producing greater compression than Fast Infoset is not very meaningful as such, because such greater compression surely comes at the expense of something else. For example, there may be a cost in extra CPU cycles, or complexity of implementation, or stronger coupling between senders and recipients due to the need to share the schemas. Such stronger coupling may be problematic in many applications involving multiple loosely-connected systems.

Therefore, although it's true that EXI can achieve greater compression than Fast Infoset, many users will not be able to actually use EXI in such a way to enjoy that extra compression. In contrast, Fast Infoset is simple. It was designed to balance compression, speed, and ease of implementation. Moreover, it is a robust and proven international standard that has been
available for about three years.



We hear that the very idea of XML compression is heresy in some circles. The whole point of XML is that the data is plainly readable, so rendering it binary takes away its core strength. On the other hand, ASCI is an expensive way to carry data across the wire. In any case, it's always good to get more info on the topic.

Posted by Joab Jackson on Feb 08, 2008 at 9:39 AM


inside gcn

  • cybersecurity (vs148/Shutterstock.com)

    NIST lays groundwork for encrypting IoT devices

Reader Comments

Tue, Oct 28, 2008 safepage safepage

Hi Joab,Liquid Technologies (www.liquid-technologies.com) took part in the Sun Microsystems backed Fast Infoset interoperability tests a while back, here is there press release from March 7 2007...Liquid Technologies Ltd is pleased to announce the successful interoperability testing of its Fast Infoset implementation with those from Noemax Technologies Ltd, OSS Nokalva Inc, Sun Microsystems Inc, and TMax Soft Inc. Fast Infoset is an ITU-T and ISO defined standard that specifies a binary encoding for W3C XML Information Set. Unlike other XML representations, the Fast Infoset standard has the dual benefits of both compression and performance, making it the ideal choice for moving large XML data between disparate low bandwidth systems or for high performance systems such as those utilising Web services. The interoperability testing effort shows that Fast Infoset is a mature, reliable technology. Software developers can now invest in Fast Infoset technology with the confidence that third parties will be able to interact with the XML data produced using a variety of vendors' products across many platforms such as Microsoft Windows, Linux and Sun Microsystems Solaris, using C++, Java or Microsoft .Net....I'd suggest this shows that the Fast Infoset movement is mature and should be pushed by W3C as a standard way of compressing XML.

Sat, Feb 9, 2008 John Schneider WA

Alessandro statements about W3C's Efficient XML Interchange (EXI) are very misleading and need to be addressed. In the interest of full disclosure, I work for http://w DOT ww.agiledelta.com the company that developed Efficient XML and contributed it to the W3C. I am also co-editor of the http://www.w3.org/TR/exi/ DOT . Alessandro states that "producing greater compression than Fast Infoset is not very meaningful..." I think the people that need high compression might beg to differ here, especially government uses that often have very limited bandwidth. Efficient XML is a quantum leap ahead of Fast Infoset in this area. Efficient XML always achieves better compression than Fast Infoset and in some cases achieves compression ratios over 10 *times* better. See http://www.agiledelta.com/p DOT roduct_efx_fi_comp.html for a summary of the W3C's benchmarks for Efficient XML and Fast Infoset compression.Alessandro also claims that "such greater compression surely comes at the expense of something else." The W3C also conducted extensive benchmarking of encode speeds and decode speeds, finding no performance penalty for the increased compression achieved by Efficient XML. In fact, the high-performance Efficient XML processor we provided (see http://www.agiledelta.com/p DOT roduct_efx.html proved to be one of the fastest processors they tested. For a head-to-head comparison of both compression ratios and processing times achieved by Efficient XML and Fast Infoset on some common department of defense messages (NCES SOAP, CoT, VMF, TDL) see http://www.agiledelta.com/e DOT fx_perffeatures.html. Alessandro also mentions "stronger coupling between senders and recipients due to the need to share the schemas" as one possible downside. Efficient XML has the ability to leverage your existing schemas to increase performance, but it does not require them and it always beats Fast Infoset whether you use them or not (see summary of W3C results above).His statement that "many users will not be able to actually use EXI in such a way to enjoy the extra compression" is false. Efficient XML Interchange can be used anywhere Fast Infoset can currently be used. BTW: The development of Efficient XML was sponsored by the U.S. government's Small Business Innovation Research (SBIR) program to specifically address the needs of the most demanding government applications. The DoD has used Efficient XML to rapidly get critical information to/from aircraft, vehicles, SOF units, satellite systems, and a variety of C2 systems. If there's interest, I'd be happy to share more information about the DoD's use of Efficient XML with the GCN community. All the best!, John Schneider CTO, AgileDelta, Inc. Editor, W3C EXI standard

Please post your comments here. Comments are moderated, so they may not appear immediately after submitting. We will not post comments that we consider abusive or off-topic.

Please type the letters/numbers you see above

resources

HTML - No Current Item Deck
  • Transforming Constituent Services with Business Process Management
  • Improving Performance in Hybrid Clouds
  • Data Center Consolidation & Energy Efficiency in Federal Facilities