Greetings

I analysed some emails in testconnectivity.microsoft.com and get to know that the some of the emails taking 5-10 seconds for delivery.

I have 2 Mailbox servers and 2 CAS on Exchange 2013 CU7.

Almost all of the delay is happening while submitting the email by Mailbox1 to Mailbox2.

Kindly suggest me how to decrease the Email delivery time.

And what factors this Email delivery time depends. 

Thanks in advance

• Edited by K 2 Sunday, July 05, 2015 8:06 PM

Hi,

Does it happen from Mailbox server 2 to 1? It sound like your receive connectors has the tarpit attribute set to default 5sec. check it by going into EMS and type get-receiveconnector "<servername>\default <servername>" | select tarpitinterval. You should see 00:00:05. You can check this to anything you want but it will depending on your requirements and bandwidth. 

• Proposed as answer by Allen_WangJFMicrosoft contingent staff, Moderator 56 minutes ago

Hi,

Does it happen from Mailbox server 2 to 1? It sound like your receive connectors has the tarpit attribute set to default 5sec. check it by going into EMS and type get-receiveconnector "<servername>\default <servername>" | select tarpitinterval. You should see 00:00:05. You can check this to anything you want but it will depending on your requirements and bandwidth. 

• Proposed as answer by Allen_WangJFMicrosoft contingent staff, Moderator Tuesday, July 07, 2015 6:29 AM

Is it happening for internal mails or external mails or both?

Hi,

As additional, we can use message trace to find out what happened about this delay deliver.
More details please refer to "Why is my message taking so long to arrive to its destination? Where is it in the pipeline? ": https://technet.microsoft.com/en-us/library/jj200741(v=exchg.150).aspx#BKMB_Whyismymessagetakingsolong

I have noticed this latency (approx 5-10 sec) in some of the internal emails irrespective of attachment size.

• Edited by K 2 Tuesday, July 07, 2015 7:53 PM

Hi,

Do you have try the suggestion as Michael mentioned?

Thanks Michael for your suggestions.

As u have mentioned, the Attribute value "TarpitInterval" on both the Exchange 2013 Mailbox Servers is 5 (Default value).

However, Why this attribute is triggered for internal emails from the User Mailbox(s) and Internal Applications.

Is the local time on Mailbox1 and Mailbox2 the same, or is it skewed by a couple of seconds ? Since each mail hop independently adds his own time stamp when appending to the header, it might be a reason.

Secondly, check the protocol logs to see exactly if there's any reason being mentioned, particulary if tarpitting is  involved. In my experience, it's only in the receive protocol logs where tarpitting is flagged (I'm not 100% sure nor do I have a link to an article). Even if the mail traffic is internal, by default the "Default <MailboxServer>" receive connector will be used on the servers hosting the recipient's mailbox. On a default installation, these logs will be in %ExchangeInstallPath%TransportRoles\Logs\Hub\ProtocolLog\SmtpReceive, provided the logging level is set to "Verbose". If it's not, you can set it on both your servers with:

Set-ReceiveConnector "Mailbox1\Default Mailbox1" -ProtocolLoggingLevel Verbose
Set-ReceiveConnector "Mailbox2\Default Mailbox2" -ProtocolLoggingLevel Verbose

You can also check the equivalent send logs on the servers originating the messages. Since it's internal emails we're after, the logging level on the built-in, by default hidden, intra-org send connector needs to be turned up to Verbose, if not already. As specified by this article, this can be done by running the following 2 cmdlets from an Exchange Management Shell:

Set-TransportService Mailbox1 -IntraOrgConnectorProtocolLoggingLevel Verbose
Set-TransportService Mailbox2 -IntraOrgConnectorProtocolLoggingLevel Verbose

Then try sending another email, and take a look in the send protocol logs on the mailbox server that originated the message. By default these are stored in %ExchangeInstallPath%TransportRoles\Logs\Mailbox\ProtocolLog\SmtpSend as specified here.