none
Need help understanding this "Product Behavior" RRS feed

  • Question

  • In this document: http://msdn.microsoft.com/en-us/library/ee157980(EXCHG.80).aspx is the following:

    <70> Section 2.2.1.19.2.14: In Exchange 2007, the server sends Sync response messages containing the <MoreAvailable> element and between zero (0) and <WindowSize> schema changes when it encounters elements external to the protocol. If the client receives multiple Sync responses that contain the <MoreAvailable> element and fewer changes than requested by the <WindowSize> value included in the Syncrequest, then the client SHOULD continue to send Sync requests to ensure that all in-protocol schema changes have been received by the client. If this Sync request and response loop is affecting network performance and synchronizing the client is of less importance than network performance, then the client SHOULD stop sending Sync requests.

    Could someone please explain:

    1) What are these elements "external to the protocol" that cause this behavior?

    2) The implication is that the SHOULD behavior (continuing sending requests) might continue indefinitely, as it might be "affecting network performance".  Can you explain this further?

    Overall, this is a rather peculiar "behavior".  Can you shed some light on best practices in dealing with it?

    Thanks,

    Guy

    Wednesday, May 19, 2010 4:58 AM

Answers

  • Hi Googley Guy,

     

    Thank you for your question.

     

    1)      This refers to any MAPI properties that do not map into the Schema for the particular type as defined by the EAS protocol.

    2)      The looping should not continue indefinitely unless you have some other issue/bug.  However, if there are many out-of-schema changes there could be a large number of Sync requests needed to ensure all in-protocol schema changes are received/processed by the client (number of changes / window size).  If there are many Sync requests/responses then that can have an impact on network performance in an extreme case.  Therefore, if bandwidth is more important than up-to-date items, you may want to stop sending Sync requests with the understanding that you may not have all of the valid changes.

     

    Regards,

    Mark Miller

    Escalation Engineer

    US-CSS DSC PROTOCOL TEAM

    Friday, May 28, 2010 6:38 PM

All replies

  • Hi Googley Guy:

    I have alerted protocol documentation team about your inquiry. A member of the team will be in touch soon.


    Regards, Obaid Farooqi
    Wednesday, May 19, 2010 3:46 PM
    Owner
  • Hi Googley Guy,

    I will investigate this issue and respond to you on this forum asap.

    Regards,

    Mark Miller

    Escalation Engineer

    US-CSS DSC PROTOCOL TEAM

    Wednesday, May 19, 2010 8:21 PM
  • That would be great!
    Wednesday, May 19, 2010 9:40 PM
  • Hi Googley Guy,

     

    Thank you for your question.

     

    1)      This refers to any MAPI properties that do not map into the Schema for the particular type as defined by the EAS protocol.

    2)      The looping should not continue indefinitely unless you have some other issue/bug.  However, if there are many out-of-schema changes there could be a large number of Sync requests needed to ensure all in-protocol schema changes are received/processed by the client (number of changes / window size).  If there are many Sync requests/responses then that can have an impact on network performance in an extreme case.  Therefore, if bandwidth is more important than up-to-date items, you may want to stop sending Sync requests with the understanding that you may not have all of the valid changes.

     

    Regards,

    Mark Miller

    Escalation Engineer

    US-CSS DSC PROTOCOL TEAM

    Friday, May 28, 2010 6:38 PM