When to have a Lync Trusted Application Pool for Exchange OWA IM Integration?

This post describes how to configure Exchange 2013 OWA IM Integration to Lync 2013. It mentions that you need to create a trusted application pool and trusted application in the Lync 2013 topology representing the Exchange 2013 MBX server and OWA. Let me try to explain why this is so and when you have to do it.

Internally OCS and Lync uses the concept of Known Servers and it maintain a list of them, the so-called Known Servers Table (KST). A server needs to be in the KST, before Lync will communicate with it. The KST is created from information found in the Lync topology. You can also add to the table by creating Trusted Application Pools from Lync Server Management Shell or from adding Trusted application servers in Lync Server 2013 Topology Builder. A Known Server has a version number attached to it. The version number is related to the Lync version used when creating the known server: 6 is Lync 2013, 5 is Lync 2010 and 0 is any.

When RTCSRV starts, it will log event 14492 from LS Protocol Stack listing all the known servers. It will also log event 14493 from LS Protocol Stack, when a server is added to the KST, and event 14495, when a server is removed from the KST.

When the Exchange UM Integration was introduced with OCS 2007, we wanted to automate the creation of known servers for Exchange UM servers. Consequently, the Exchange UM Routing component adds the Exchange servers it finds to the KST. The servers it finds are the Exchange servers hosting SipName UM Dial Plans. It logs event 44024 from LS Exchange Unified Messaging Routing at startup listing the servers it found. It will also log events while running; indicating added or removed Exchange servers. When adding Exchange servers to the KST it sets the version to 0.

That was the background. Let us look at what happens.

We will assume an environment with Lync 2013 and an Exchange 2013 server called ex1.contoso.com. It is running all Exchange 2013 server roles, is hosting a SipName UM dial plan and Exchange OWA IM Integration has been configured.

When the Exchange UM Routing component starts it will add the entry {fqdn=ex1.contoso.com, version=0} to the KST. At this point Exchange UM integration and Exchange OWA IM Integration to Lync 2013 both works.

The Lync 2013 administrator then adds Trusted Application Pool and Application representing the Exchange server and OWA to the Lync topology. This means adding the entry {fqdn=ex1.contoso.com, version=6} to the KST. However, it is not legal to have two entries for the same server with different versions. The event 14563 from LS Protocol Stack is logged and both entries are removed from the KST.

This has the effect that Exchange UM and OWA IM Integration stops working, since there is no entry for ex1.contoso.com in the KST.

How to get Exchange UM and OWA IM integration working again?

Option 1

Remove the Trusted Application Pool and Application representing the Exchange server and the integration will start working again shortly.

Note: The caveat with using this option is that you have to remember to add the Trusted Application Pool and Application back to Lync 2013, if the Exchange server stop hosting a SipName UM dial plan.

Option 2

Restart RTCSRV on the Lync 2013 server. The restart will still produce the error event 14563 and a warning event 44030, but afterwards you will see an Information event 14493 from LS Protocol Stack adding the Exchange 2013 server to the KST. At this point, the integration works again.