Mike Stacy's Blog

Lync/OCS Coexistence and Static Routes

Having worked heavily with static routing since the pre-release versions of Lync Server 2010 I’ve discovered that while the cmdlet documentation is pretty good on how to create and manage static routes, there is no information on managing static routes in a mixed environment containing both OCS and Lync Server.  Recently I had a discussion on this topic with some folks at Microsoft and was able to get some helpful information on the topic.

If you’ve created routes in both OCS and Lync Server you’ve no doubt realized that the process is pretty different.  As I referenced in my original post on creating static routes in Lync, Lync enables administrators to create routes with different scopes.  Specifically, you can create routes that apply either at the global or registrar level.

Something else apparent as you work in a mixed environment is that Lync is able to use the static routes that have previously been created in OCS but the reverse is not true.  Further, a new route added in OCS cannot be leveraged by Lync, and creating the same route in both places leads to routing errors on the OCS side.

The reason for all of this is because of the new mechanisms in Lync for storing and managing static routes.  When the Merge Topology process is executed in Topology Builder, the existing OCS routes are copied into the Lync configuration and are therefore able to be used by Lync.  New routes created in Lync are only stored in the configuration in such a way as to be accessed by Lync, and since there is effectively no synchronization of configuration from Lync to OCS these routes are not able to be leveraged by OCS.

Therefore, during the period of coexistence the proper way to create routes that you wish to be accessed by both OCS and Lync is to create them in OCS and re-run the Merge Topology process.

Great Tool - ‘Find Lync Versions’

I’d love to give credit by name but I only have the name “stumper” to thank for a great OCS/Lync tool.  Those of you who spent much time managing OCS environments are probably familiar with the Client Version Summary that was easily accessible in the OCS admin tool.  Unfortunately the only way to gather equivalent information in Lync was to write the appropriate SQL queries to do so.  However, our friend stumper has written a nice, simple app to get this information.  You can even get the source code if you want to make changes.  Here’s a link:  http://www.stumper66.com/software/lync.html

Here’s an image of it from stumper’s site:

 

The summary tab is very handy to get a quick look at your client usage.  In my case it was great to see how many users were still using MOC instead of Lync.

ICE/STUN/TURN with Polycom HDX and Office Communications Server 2007 R2

There is an interesting OCS-related update coming to the Polycom HDX line of videoconference systems with the 2.6.1 release scheduled this quarter.  First, a quick overview in case you’re not familiar with the HDX product set.  The most common use case for an HDX is as a room-based video conferencing system, such as the scenario depicted in the image below.

There are also other HDX systems designed for personal use and special use cases, like medical carts:

HDX 4000	HDX Medical Cart

   

There are a number of characteristics that all HDX systems share.  First, they can all act as native OCS endpoints.  Specifically they can register to OCS, publish presence, show an OCS contact list (including presence and groups), and make/receive video (and audio) calls to other OCS clients.  Second, all HDX systems are high definition (at least 720p and up to 1080p depending on the model).  This high definition experience includes OCS-based calling scenarios between HDXs, meaning that I can call from my personal HDX 4000 at home to a conference room in one of our offices and receive an high definition experience, even though OCS doesn’t have native support the for the H.264 codec being used in the call.

All that being said, the HDX is already a great complement to a Microsoft UC environment.  However, Polycom is strengthening the capability of the HDX line by adding OCS-compatible ICE/STUN/TURN to the software.

If you’re not already familiar with ICE, STUN, or TURN, there are plenty of resources on the internet that go into detail about these protocols.  The quick overview is that these methodologies are used in conjunction with SIP to enable media (such as audio and video) to traverse firewall/NAT devices.  If you’ve ever participated in a call to or from a Communicator client that was logged on via an OCS Edge server (from the internet without VPN), then you’ve used these protocols.

So what does adding ICE/STUN/TURN to HDX enable?  First, it extends HDX calling scenarios to remote users rather than just Communicator clients within the organization or connected over VPN.  Second, it enables HDXs outside the organization to register to OCS and participate in audio/video calls just like a Communicator client.  Usually these external systems will be HDX 4000s, but it would also be useful for organizations that have internet-connected branch offices or in the medical industry when a consult with an external physician is required.  Third, it enables federated scenarios whereby calls can be made between HDXs or MOC/HDX in separate organizations.

Because this connectivity only relies on the OCS Edge server, no additional infrastructure is required to support this capability.  This is consistent with the rest of Polycom’s OCS integration strategy, which leverages customers’ existing OCS architecture rather than requiring additional gateways or other hardware.

Stay tuned for periodic Microsoft UC updates from Polycom in the quarters ahead – the integration between these two is only going to get stronger.

OCS 2007 R2 is now supported on Windows 2008 R2

All the details for support are in KB 982021:  http://support.microsoft.com/kb/982021/.

Additionally, KB 982020 outlines issues when upgrading the domain to Server 2008 R2 and the way to fix these issues by re-running forestprep.

The Unified Communications Community on Twitter

Like many techies out there, I’m usually very quick to sign up for new services/web sites/tools/etc. if even for no other reason that to see what it is and how it works.  For me, Twitter was no exception.  When it first debuted I signed up for an account and, shortly after, summarily decided that it was worthless to me.  This is quite different than my viewpoint on technologies like blogs, wikis and other now-commonplace tools that we all use with great benefit today.  I simply thought that I was not in the target demographic for Twitter – I wasn’t going to post my random thoughts on any sort of consistent basis and even if I did, the only people that might care weren’t really the type who would be so connected as to benefit from this method of communication.

However, a few weeks ago I a search result I received was linked to a Twitter post, and furthermore I knew the person who had posted it.  A quick look at his Twitter page indicated an active discussion on topics that are near and dear to my heart, namely Unified Communications.  This put me on a bit of a mission to see how many people in the community were using Twitter and what information was being communicated.  I looked at the active topics and followed links to those users’ pages for an hour or so, more out of interest in the topics than anything else.  Based on this I decided that I was missing out and needed to join in.  It’s impractical to use Twitter via a web browser for more than a few tasks (too many clicks, searches, etc.) so I immediately downloaded TweetDeck, which allows you to have search columns that act similar to RSS feeds.  It also includes columns that show your direct messages and mentions.

I quickly added search columns for #ocs and #ucoms and right away I had great views of all the discussions on these topics, complete with toast-style notifications of updates.  Tweets on these hashtags include (but are not limited to) things like Q&As, rants about an issue, press releases, and companies/recruiters looking to hire.

Like anything, the system is not without its downsides.  Using a nice tool like TweetDeck makes ReTweeting annoying since all the RTs contain the same hashtags as the original Tweet.  Also, a limitation specific to TweetDeck is that there is no quick option off the taskbar icon to suspend the toast notifications.  I do a lot of presenting so this means I have to either so modify the application settings or exit the app altogether every time I present.  I usually end up exiting since it’s fastest and I have a tendency to forget until I get the first toast during the presentation.  Respecting that the system is in a full screen presentation and automatically modifying the behavior of the app would be ideal, but the quick menu option would be a step in the right direction.

I’m still not the type of person who is particularly fond of (or thinks other people have an interest in) posting my random thoughts, so I’m more of a lurker than anything else.  Even so, I enjoy the steady stream of data complete with plenty of bit.ly links if something of interest pops up.  If you haven’t checked out Twitter recently for use in a business context then I would recommend revisiting it.

Creating OCS Address Book Entries for Polycom RMX Meeting Rooms

In case you’re not familiar with the Polycom RMX, it’s a platform of audio/video MCUs that can integrate with everything from voice PRIs to legacy video devices to the newest 1080p endpoints.  Starting with version 4.1 (current version as of this writing is 5.0), it also supports integration with OCS.  Specifically, OCS users can call into the RMX via audio or video by simply making a Communicator call to an RMX meeting room or entry point.  The RMX can bridge those MOC users with other non-OCS SIP endpoints as well as H.323 endpoints.  For more information on the RMX check out Polycom’s website.  While it’s fairly easy to call a meeting room from Communicator by dialing it manually, for folks that use MOC in conjunction with RMX meeting rooms consistently it is desirable to have quick dial access to the rooms.  An easy way to do this is using contact objects in Active Directory.

First, create a regular contact object in Active Directory.  You can enter any information you’d like in the required fields, just be sure to make the Display Name match what you’d like your users to see in the OCS address book.

Next, use adsiedit to assign a SIP address to the contact.  If you do not have adsiedit installed, just download the appropriate Windows Support Tools for the system you are using.  Once they are installed you can launch adsiedit via Start –> Run –> adsiedit.msc:

Next, navigate the Domain context to the location of your contact object and open its properties.  Scroll down to msRTCSIP-PrimaryUserAddress.  You’ll notice that it currently has no value set:

Enter the address in the format sip:<MeetingRoomID>@<RMXRoutingAddress>.  For example: sip:1000@rmx.domain.com.

Because you’ve defined a SIP address for the contact, OCS will automatically add it to the address book.  However, be aware that the address book only synchronizes every 24 hours at 1:30am local time (by default).  You can force the address book to regenerate faster using the abserver –syncnow command.  Be sure to allow time for AD replication before you run the address book synchronization.  For more details on the address book see http://technet.microsoft.com/en-us/library/ee323612(office.13).aspx.

Once the address book has synchronized and your client has downloaded the updated address book file, you will be able to find the room by name in MOC.  Since the RMX is not registering to OCS the presence will show as unknown, but assuming that your environment is configured correctly you will be able to make audio and video calls to that contact from the address book just like you can by manually typing in the SIP address.

Quarterly UC Virtual User Group meeting tomorrow

This is a reminder that the Microsoft Unified Communications Virtual User Group will be holding it's last quarterly meeting of 2009 tomorrow at 2:00 PM Eastern (-5 GMT). We've had a great year getting UCVUG off the ground and look forward to ending the year with an awesome presentation by Chris Norman from Microsoft. Chris will be presenting on Response Groups in OCS 2007 R2.  You can register for the event at http://ucvugdec2009.eventbrite.com/.  The Live Meeting link is also posted on the UCVUG website.

We will also be testing our new chat room on the UCVUG website prior to the meeting. Just logon to http://www.ucvug.org at 1:30 PM Eastern (-5 GMT) and click on the "Chat" tab to join us for a "meet and greet" chat session.

Thanks again for your continued support of UCVUG and we look forward to seeing you at the meeting tomorrow!

OCS R2 Integration with Exchange 2010 OWA

Exchange 2010 includes the ability to add presence and IM integration right into the OWA experience.  This adds a number of handy features:

- Presence for internal and federated OCS contacts

- Ability to start and maintain chat sessions directly from OWA

- OCS contact list integration, including add/remove contacts and groups

- Ability to control your presence state from OWA

Now for some screenshots…

Presence integration including context menu additions:

 

Chat window and the ability to control multiple simultaneous chat sessions:

 

Your presence and presence control:

 

Contact list integration with groups and presence (the contact list is in the left pane below your folders):

 

 

Installation and configuration was fairly straight forward, only a few items to point out.  First of all, here’s the download link and link to documentation.  Right now there’s a missing link in the documentation for a UCMA update for Windows 2008 R2.  That update is part of the 968802 package here.  All you need for this integration is to apply the UCMARedist.msp update to your Exchange 2010 server.

Now, a few things to point out in terms of the integration process.  First, make sure to put the spaces in between each pair of characters in the IMCertificateSerialNumber web.config value.  If you don’t the integration won’t work and you’ll get an error in your event log:

Log Name:      Application
Source:        MSExchange OWA
Date:          11/12/2009 11:36:28 AM
Event ID:      87
Task Category: InstantMessage
Level:         Error
Keywords:      Classic
User:          N/A
Computer:      excassrv1.corp.evangelyze.net
Description:
An exception was thrown while attempting to load the IM provider .dll file.

File: C:\Program Files\Microsoft\Exchange Server\V14\ClientAccess\OWA\bin\Microsoft.Rtc.UCWeb.dll
The local certificate specified was not found in the store for local computer.

Next, when you run the command to add the InstantMessagingType value to your OWA virtual directory (Get-OwaVirtualDirectory | Set-OwaVirtualDirectory -InstantMessagingType OCS) you will get errors for any CAS servers that are not running Exchange 2010:

Property InstantMessagingType can't be set on this object because it requires the object to have version 0.10 (14.0.100.0) or later. The object's current version is 0.1 (8.0.535.0).
    + CategoryInfo          : NotSpecified: (0:Int32) [Set-OwaVirtualDirectory], InvalidObjectOperationException
    + FullyQualifiedErrorId : 79956A7E,Microsoft.Exchange.Management.SystemConfigurationTasks.SetOwaVirtualDirectory

This can be safely ignored – the value will be set on Exchange 2010 servers.  If you want to be sure just run get-owavirtualdirectory –server <Ex2010CASSrvName> |fl where <Ex2010CASSrvName> is the name of your Exchange 2010 CAS Server.  Among the output you should see this: InstantMessagingType: Ocs

Lastly, the documentation doesn’t seem to be complete in the “Configure Office Communications Server” section.   You’ll need to add the FQDN of the certificate that you specified in the web.config as a trusted host to your OCS environment.  To do so, right click your pool and select Properties –> Front End Properties.  Click the Host Authorization tab and then click Add.  Add the FQDN of your CAS and check the “Throttle as server” and “Treat as authenticated” check boxes.

It will take a few minutes for your OCS front end(s) to pick up this change.  Depending on your level of impatience (mine is high) You can force it by restarting the front end service but this will disconnect all your users from OCS, so do this with caution if you are working in a production environment.

Lastly, in case you’re curious which user agent this connection uses, here you go:

That should do it.  The entire process took about 5 minutes and works perfectly.

Polycom HDX 4000 and RMX 2000 with OCS

I spent a good amount of time the other day working on the setup, integration, and various scenarios of Polycom’s HDX endpoints and RMX bridges with OCS.  Previously I had provided assistance to customers who needed help getting an HDX to work with OCS but I hadn’t had the opportunity to do everything from start to finish.  Overall the HDX integration was straight forward, even easy (it took about 5 minutes) while the RMX could use some help in a few areas.

First of all, a quick description in case you are not familiar with these devices.  The HDX is primarily targeted as a room-based videoconferencing system but the 4000, with its incnluded monitor, could also be used as a personal video conferencing system though the price point will probably reserve that right primarily for CxOs, videoconferencing system architects, and Polycom employees.  :)  Like the other video systems on the market the HDX does not yet support RTAudio/RTVideo and therefore, with regards to OCS, can only be used for PC to PC calling (MOC <-> HDX).  This limits the calls to CIF quality only, which is ok for use on a PC but of course doesn’t show off the HD capabilities of the HDX.

Setup was very simple – just create a user account (or use your own) and configure the HDX settings to point to OCS.  From there you can choose to get your contacts from Active Directory (using the LDAP configuration) or from your OCS contact list.  The benefit of the latter is that you get presence information from your contacts as well. It would be great to have access to search AD as well as have presence from your contacts but unfortunately there is currently no way to access both of these at the same time yet.  Another thing to consider is that the HDX doesn’t support enhanced presence, so if you create an OCS account for the conference room you need to disable enhanced presence, otherwise it will always show as offline.  Polycom’s HDX documentation outlines all this for you.

The RMX is a bridge that (among many other features) hosts audio and/or video conferences.  With the OCS integration a MOC user (including an OCS-connected HDX) can call via audio, video, or both into a conference room that exists on the RMX.  Obviously you could do this with OCS alone but the RMX extends the capability to standards-based systems so that you can have a mix of endpoints in your conference.  You can also integrate PRIs into the RMX to enable PSTN audio participation.

However, the RMX integration was not without its issues.  First, the requirements to get TLS working with OCS aren’t well documented or explained.  Through quite a bit of trial and error you can make it work but the most minor changes, like modifying certain fields to use an FQDN instead of an IP address, unexpectedly cause the integration to fail despite the apparent support of name entry.  Especially when you are dealing with TLS this is counterintuitive.  Second, the certificate upload process is quite a hassle, to say the least.  Despite the UI providing acknowledgement of certificate uploads, pfx files fail to load properly and in the end you are forced to use external utilities to break down your certificate into a combination of .pem and password files.  Even then you have to perform the upload from the web client (there is an installable desktop client as well) and apparently not from Windows 2008 or Windows 7.  I had to boot my XP mode system and connect from IE within the VM before the certificates actually took, despite receiving verification from the other upload methods.

However, once this was all sorted the process to connect to a conference was very simple.  There are also options to register conference rooms but I haven’t had the chance to try this with OCS yet.  I haven’t begun to touch on all the capabilities of the RMX in particular (check the Polycom site for that), but I hope that Polycom investigates the OIP program for the RMX 2000 as well as the upcoming RMX 4000.  It would be great for customers to have an all-in-one OCS voice gateway and video conferencing bridge option.

9 Months of OCS 2007 R2

I’m writing this from the plane on the way from Denver to NYC to spend a week with a customer of ours.  This particular customer is in the Enterprise Lighthouse program, which means that they are moving to OCS 2007 R2 as the voice platform for at least 2000 users.  In addition to this they are also migrating from Sametime to OCS for IM, presence, and web conferencing for their entire user population (150,000+ users).  It’s my second project of this scale for OCS and, from what I can gather from my records, around my 65th OCS project altogether.  Of those ~65 projects 52 included voice for at least a subset of the users.  The customers have ranged in size from 50 users to 160K users, with most falling in the middle – a bunch of companies in the 3000-20000 user range.  I was just sitting here reflecting on these projects and thought I’d capture some of these thoughts.  First, a some charts of various data points:

As you can see, we used primarily NET/Quintum gateways, with Direct SIP connections coming in at a close second.

 

Nothing surprising here, the PBX market leaders are reflected in our implementation stats as well.

 

I’m curious how our CallManager version integration breakdown matches up to Cisco’s worldwide deployment percentages.

 

Now, some less data-specific observations.  In general we saw faster voice adoption by organizations who a) were running an older PBX and b) had little to no call center staff.  This manifested itself in a few ways – expanded pilot group, addition of capacity to voice components, and/or a more accelerated move to production.  Not surprisingly smaller organizations were more likely to move towards production, which was typically because those organizations were more nimble and had a more manageable political environment.

As a consulting organization, predictability of the platform is a major consideration for us and has been a strong point of OCS to date.  It allowed us to create a standard fixed price model for almost all our services, which in turn reduced customer’s fear in moving to the platform since they knew we stood behind our work and would, if something went awry, share in their pain.  Happily for both us and our customers this did not happen – all issues/complications/etc. were properly accounted for in the engagements and covered as expected by our pricing models.

So where does everything stand today?  The obvious state of things related to voice is the lack of features as compared to traditional PBXs.  This is no secret, and will largely be addressed in the OCS Wave 14 product which we’re expecting to see in about a year.  Despite some of these feature deficiencies, there are numerous organizations for whom we’ve helped deploy OCS as a replacement to one or more PBXs.  By and large users need to be able to make/receive phone calls and subsequently forward, transfer, or conference those calls.  OCS can easily handle all of that and display presence all the while.  Reuse of existing telephony endpoints was a showstopper for a number of customers but the introduction of SmartSIP earlier this year had started to break down those deployment barriers. 

What’s missing is a better story for geographic redundancy for larger organizations.  The recent resiliency white paper attempts to address some of that but the latency requirements are too stringent for most organizations, even those with deep pockets, since global organizations will likely have datacenters that are more than 15ms apart.  Luckily I’ve been architecting around those shortfalls since the LCS days so there’s not much we can’t deal with as long as you can handle a ~15 minute outage in the event that one of your datacenters is completely destroyed.  So far that has been viewed as a reasonable recovery time.  However, I’m hoping that additional work is done to complete this story in Wave 14.

The other major missing items such as Mac/Linux support are already expected in Wave 14 so there’s not much point to hash them out here.  One thing I haven’t heard much about but would really like to see is a better story around video integration.  The lack of codec support in the MCU today is a severely limiting factor, and the recent announcement of Cisco to purchase Tandberg might make things difficult for Microsoft if Polycom is the only remaining integration partner.

Overall though one of the most difficult concepts to convey is the paradigm shift from “the phone on my desk” to integrated, presence-enabled communication.  It’s hugely powerful and Microsoft’s vision and ambition to get there is right on target.  The extensibility of the platform dramatically strengthens this vision.  Every time we bring on a new customer that does not yet have OCS you realize how much easier your communication with existing OCS-enabled internal, partner, and customer contacts is.