Magic Quadrant for Unified Communications 2007

Here is the magic Q for Unified Communication 2007. This report is generated by Bern Elliot (Gartner) dated 20 August 2007.

This is the link to the web : http://mediaproducts.gartner.com/reprints/nortel/150273.html

Guide on OCS Database tuning (SQL)

For Midrange (8 GB of memory) and High End (12 GB+ of memory)

>If 32-bit version of SQL Server is running, enable Lock Pages in Memory (http://r.office.microsoft.com/r/rlidOCS?clid=1033&p1=LockPage) and enable AWE (http://r.office.microsoft.com/r/rlidOCS?clid=1033&p1=AWEon).

>These settings are not necessary if SQL server is 64 bit.

>Add a SQL Server trace flag for T4618 (http://r.office.microsoft.com/r/rlidOCS?clid=1033&p1=kbT4618)

1) Open SQL Server Configuration Manager
2) Select SQL Server instance, right-click, and then click Properties.
3) In the Advanced tab, click Startup Parameters.
4) Append ;-T4618 to the end of the Startup Parameters list and click OK.
5) Restart SQL Server.

Communicator Address book download (Galcontacts.db)

Client "Address Book" File Download Process http://technet.microsoft.com/en-us/library/bb894482.aspx . If a client is accessing the URL for the first time, upon successful connection, the client attempts to download the current full data file. On subsequent days, the client attempts to download a delta file based on the last full synchronization date.

Typically, this delta file is based on the previous day. If the client is offline for a day or more, it determines which delta files it must download to get up to date. If the client is offline for more than 30 days, it attempts to download the full data file.

Storing this information in a local database reduces the time taken to synchronize information on the client computer with the latest information stored in Active Directory, thereby significantly improving the GAL search process. The location of this database is : Documents and Settings\\Local Settings\Application Data\Microsoft\Communicator.In the event of a download failure because of network connectivity or other issues, the client retries in time intervals that double on each failure (1 minute, 2 minutes, 4 minutes, and so on, up to a maximum of 64 minutes).

Any data that was downloaded before the failure is discarded, and the retry begins again at the beginning. If the failure persists for more than 24 hours, a warning is displayed, and an application event is added to the Event Log.

When the client logs in, it determines if it has been more than 24 hours since the last download. If so, then the current download occurs immediately. Otherwise, download is scheduled at 00:00 UTC (Universal Coordinated Time, also known as GMT).

Tracing in Office Communicator

Tracing can be enabled on Communicator by group policy or manually on the client. Enabling event logging will write communicator events into applications event logs. Protocol level logging files can be found at ..\User\Tracing in XP or Vista. These files can be read by notepad but I recommend using snooper which is in OCS resource kit.

Conversation History Settings

The settings in "Personal Information Manager" such as

-- Save my instant message conversations in the Outlook Conversation History folder

-- Save my call logs in the outlook Conversation History folderare being published to the backend database, rtc.PublishedStaticInstance as binary, by the client.

<otheroptions>
<permissions> <personalinformationmanager>outlook</personalinformationmanager> <autoretrieveooffromoutlook>false</autoretrieveooffromoutlook> <publishcalendarpresence>false</publishcalendarpresence> <imautoarchiving>false</imautoarchiving> <calllogautoarchiving>false</calllogautoarchiving> <publishmeetingsubjectandlocation>false</publishmeetingsubjectandlocation> </permissions> </otheroptions> </publication> </publications> </publish>

Using Hardware Load Balancer for OCS 2007 Front-end Server

To increase the scalability and higher redundancy for OCS Front-end server, we have to use hardware load balancer. Below is some of the support hardware offerings from F5, Cisco, Foundry, Nortel, and WebMux.

http://www.f5.com/solutions/deployment/index_lcs.html

http://www.f5.com/pdf/deployment-guides/microsoft-ocs-ltm94-dg.pdf

http://www.foundrynet.com/pdf/wp-ms-office-comm-serv-2007-load-bal.pdf

http://www.cainetworks.com/papers/webmux/lcs2005.htm

http://blogs.technet.com/uc/archive/2007/03/30/load-balancers-with-live-communications-server-2005.aspx

http://office.microsoft.com/en-us/communicationsserver/FX011526591033.aspx

Disaster Recover for OCS 2007

As we making OCS 2007 as the main call manager, high avaiability and scalability would one of be the main components we (IT person) to consider. Scalability in OCS 2007 would be easy as Microsoft has addressed this metter. The second point which other players in the IP Telephony industry is questioning Microsoft is; can OCS 2007 have the disaster recovery solution?

There are some solutions from others Microsoft team proposed which using backup and recovery method on the OCS 2007 server. Can it be automated? This has makes me to investigate more in my current lab. I've trying to create a auto service failover to DR server when the production server is down. I'll continue more after my lab test.

I've browse around, M Senthil Kumar blog has a discussion on backup OCS 2007 configuration guide. (http://blogs.msdn.com/mskumar/archive/2008/02/11/backing-up-ocs-2007-global-config.aspx)