Currently Browsing: Exchange UM

Exchange 2010 UM cannot find Lync 2013 Edge pool

On my initial Lync 2013 project I encountered a problem with the Exchange 2010 UM server not finding a path to the Lync 2013 Edge pool. I was receiving the MSExchange Unified Messaging event 1438, as shown below.

I was able to resolve the issue by running the following PowerShell command on the Exchange UM server (replacing lync-edge.domain.local with your Lync Edge pool’s FQDN):
Get-UmServer | Set-UmServer -SIPAccessService lync-edge.domain.local:5062



Log Name:      Application
Source:        MSExchange Unified Messaging
Date:          11/14/2012 7:43:00 AM
Event ID:      1438
Task Category: UMCore
Level:         Warning
Keywords:      Classic
User:          N/A

The UM server has been configured to automatically use the Communications Server Audio/Video Edge resources associated with ‘lync.domain.local:5061’. Inbound and outbound calls involving remote users (located outside the enterprise) might be failing using the current Communications Server Audio/Video Edge resources. To correct the issue, set the SIPAccessService property using the Set-UMServer cmdlet. The Microsoft Exchange Unified Messaging  service will start using the Communications Server Audio/Video Edge resources corresponding to the new value.


Phonetic Names

For some people the spelling of their names doesn’t match the enunciation. It may bother these people when they hear their name enunciated incorrectly on conference entry/exit announcements, Outlook Voice Access prompts and outgoing message. The Text-to-Speech capabilities in Lync and Exchange UM are very good, but not perfect.

So, how can we help these people? We can force the text-to-speech engine to better match the actual name. Each user object in Active Directory (in Windows Server 2008 and later schemas) has attributes that can override the text-to-speech with phonetic spellings of the name. These are msDS-PhoneticDisplayName, ms-DS-PhoneticFirstName, msDS-PhoneticLastName, msDS-PhoneticDepartmentName and msDS-PhoneticCompanyName. By spelling the names as they sound and placing these spellings in one or more of these attributes will cause the text-to-speech engine to enunciate the names as the user wishes. This could also be used to override a full or foreign name with a shortened or localized name, such that Steven can become Steve.


As an example, one of my co-worker’s first name is Sian. The name should sound like the colour cyan. So, I added cyan phonetic spelling to the msDS-PhoneticFirstName attribute in her Active Directory user object. Now, whenever she is using our company’s Lync and Exchange UM she hears her same as expected. The only fault with this is that we cannot control this in federated connections. When she joins a Lync conference hosted by a partner organization, it won’t have the phonetic entries and she hears she-an instead of cyan.