IT06447: DFDL PARSING PERFORMANCE FOR LARGE MODELS DEGRADES AND PLATEAUS AS MORE MESSAGES ARE PARSED

APAR status

• Closed as program error.

Error description

• For a DFDL schema that contains a large number of elements
  (i.e. thousands of elements or more), and that defines a number
  of different messages, the performance of the parser when
  parsing an individual message will degrade as each possible
  message type is parsed. The performance of the parser will
  become stable once the parser has parsed each possible message
  type.
  

Local fix

• NA
  

Problem summary

• ****************************************************************
  USERS AFFECTED:
  All users of IBM Integration Bus V9.0 who use DFDL message
  models with thousands of elements, where multiple different
  messages are processed. This includes the DFDL SWIFT model and
  other very large DFDL industry format models.
  
  
  Platforms affected:
  MultiPlatform, z/OS
  
  ****************************************************************
  PROBLEM DESCRIPTION:
  When using DFDL to parse a model that contains a large number of
  elements (i.e. thousands of elements or more), the parser
  maintains a set of internal caches that use hash maps that are
  keyed on the element name. These hash maps do not resize and
  have an initial size that in inappropriate for large message
  models. This means that for large models, the hashed keys start
  to collide and any search of the hash map degrades in
  performance.
  
  For a large message schema that contains a lot of elements, as
  more different messages are parsed, the internal caches fill up,
  so performance is seen to decline and then plateau once every
  type of message has been parsed.
  
  There are a number of resource name changes between WebSphere
  Message Broker and IBM Integration Bus Version 9.0. For details
  visit
  http://pic.dhe.ibm.com/infocenter/wmbhelp/v9r0m0/topic/com.ibm.e
  tools.mft.doc/bb23814_.htm
  

Problem conclusion

• The product has been modified so that the internal caches are
  resized to maintain efficient lookups and prevent the
  performance degradation.
  
  ---------------------------------------------------------------
  The fix is targeted for delivery in the following PTFs:
  
  Version    Maintenance Level
  v9.0       9.0.0.4
  
  The latest available maintenance can be obtained from:
  http://www-01.ibm.com/support/docview.wss?rs=849&uid=swg27006041
  
  If the maintenance level is not yet available,information on
  its planned availability can be found on:
  http://www-1.ibm.com/support/docview.wss?rs=849&uid=swg27006308
  ---------------------------------------------------------------
  

Temporary fix

Comments

APAR Information

• APAR is sysrouted FROM one or more of the following:

  IT06329

• APAR is sysrouted TO one or more of the following:

Fix information

• Fixed component name

  INTEGRATION BUS

• Fixed component ID

  5724J0530

Applicable component levels

• R900 PSY

     UP