So you want to receive faxes in your email. And because exchange 2007 says it will work you want to make it happen. Wow the neat thing is exchange will handle T.38 inbound faxing real well. The problem is how do you make it work. There are a lot of discussions out there about this. And it does work well. The big issues with this is how the fax call makes it to exchange UM.

So generally there are 3 ways faxes can be setup for a user.

1. user has a separate fax number going or associated with his mailbox.
2. User uses the same number as his extension
3. Fax number goes to a resource mailbox and it is then distributed or viewed by other  users by attaching and viewing the mailbox.

Number 1 is probably the most common one and most widely used. Because number 2 the users phone rings and they pick it up and have to transfer it to their voice mail when they hear the fax tone. (this is cumbersome at best)

So generally when you setup number 1 you need to be sure that inband signaling is enabled on exchange. In SP1 it is turned off by default. to enable this find the globcfg.xml file located in  \Program Files\Microsoft\Exchange\bin Change the EnabledInbandFaxDetection setting to true and restart the UM services.

Once this is complete you should be able to send fax calls to exchange and it should detect the FAX tones.

there may be some PBX configuration that is needed it just depends on the PBX.

Hope this helps

So is time important in a Microsoft active directory Domain? Well we all know the answer to that. YES it is Kerberos must have time with in 5 min. most people that deal with active directory know this. And for most people when the create the Domain time just works. So we often take if for granted. So why I am blogging about time on the office communications server 2007 blog? Well because time is important to many more things then Kerberos. Now I probably will not reference OCS in this post but just realize time is important.

I tend to think that time should be one of the first parts of a corporations Active directory design. Why you ask? Well have you ever worked anywhere that the time on phone on  your desktop and the time on your computer were different? how about trouble shooting. When I trouble shoot I often pull logs from many different places. I even pull logs from wireshark using my computer. and I may not have time to review them for a day or two. I am often looking for time correlation of events. but if the time is off from my router to my switch to my phone system to my pc how can I correlate that the event happened at a certain time and know what effect it had on other machines or equipment.

How about companies that bill by the minute. have you ever seen a person take a call and take notes on their computer then the time on the notes be off from the time on the phone. This could cause many problems with billing and other things. So it is critical to plan time into your entire system. This is one of those areas you will here me complain about the Separation from the network team and the Active directory team. Both teams need to be consistent and pointing to the same thing across the enterprise. but often they do not talk and often they are pointed to different time servers or in some cases not even using a reliable time source.

So how should it look? you ask well in my opinion your core router should point to a time source on Internet preferably a resource pool of time servers. Then your PDC can point to the core router or point to the same pool of time servers. either will work. all the other domain controllers should be using domain Hierarchy. Remember if you move your PDC then you need to validate that time is pointed to the right place. using w32tm /monitor on a single DC will tell you what all the DC's are looking at for time. and will tell you what the PDC is looking at for time.

I have seen many environments that use the default and because the default time server is not available for what ever reason it reverts to the LOCAL time source or BIOS clock of the PDC. I am sorry but the CMOS clock to me is unreliable I have seen to many that were not correct or the battery's were dead and every time you lost power the PDC would start changing time on the rest of the machines causing major problems.

here is a great source of time information i have included 2 links one to ntp pools which is a good ntp source. and the other is to the Microsoft Technet article. .

http://www.microsoft.com/technet/prodtechnol/windowsserver2003/technologies/security/ws03mngd/26_s3wts.mspx

http://www.pool.ntp.org/zone/us

hope this helps someone

WOW I had an epiphany today. I have attended a couple of presentations by Allen Shen formerly of Microsoft. And some presentations by Laura Chappell and both of them inspired me to really look at what is happening with Calls and Call control for OCS. Allen tried to explain ICE, STUN, and TURN to us in a presentation at Tech Ed. Which was really good but I still had some questions and still was not sure about it.

If any of you attended this session or have seen one of allens sessions on ICE, STUN or TURN. then many of you realize it starts out very well but there is a point in the presentation where you become a little lost Generally this is when the lines start to cross. it is the same if you start to read the RFC's on those protocols as well. So I started last night trying to explain a little about the AV edge. And in the diagram it shows 2 users with voice traffic going through the AV edge server.

Well I wanted to see more about this so I could confirm some things for myself. I began sniffing traffic (Thanks Laura Chappell you created a monster my wife can't understand why I am going nuts wanting to get home to see what the sniffs look like) to see what is actually happening. So here is a synopsis of what I found and how I found it.

Let me explain what and how I tested this so far I will do more over the next couple of days. (it is so cool and so much fun to see how it all works)

OK

user A is at home behind a nat firewall with in internal IP of 192.168.33.74 the Router has a Public IP of 76.34.37.65

User B is on a wireless aircard from sprint. with an ip of 70.209.48.141 

OCS AV edge server for the user A OCS environment is 74.88.32.5

User B is a a member of another federated company so they are not members of the same company or the same OCS systems.

User B is calling user A

User B right clicks on the users and says make communicator call to user A.

user B begins sending sip traffic to his OCS server. I am not going to go through all this but ultimately the 2 access edges pass sip call control traffic back and fourth to each other at some point USER A sends back a list of possible IP's that user B may be able to contact him at. here is the list.

a=candidate:ji0XDiGUsQde1Fr4OZC9yqAwEIZBMAUavSjGXQkOfos 1 YPqc4I+bY14FUpKJh2Rtlg UDP 0.830 192.168.33.74 25856

a=candidate:ji0XDiGUsQde1Fr4OZC9yqAwEIZBMAUavSjGXQkOfos 2 YPqc4I+bY14FUpKJh2Rtlg UDP 0.830 192.168.33.74 56576

a=candidate:LRJGqNyquostNMw5jcDIBsW31GtYQ/cYw8fY2cWZ/Ek 1 V8p3r9V8NrclMS0ZNshlbA TCP 0.190 74.88.32.5 59722

a=candidate:LRJGqNyquostNMw5jcDIBsW31GtYQ/cYw8fY2cWZ/Ek 2 V8p3r9V8NrclMS0ZNshlbA TCP 0.190 74.88.32.5 59722

a=candidate:iIp3PUfo9f8PUICapk9cbvNixh3G//E0YcN/H1BS2uU 1 gJOKyEAWqEO3fRvwo7zibw UDP 0.490 74.88.32.5 52410

a=candidate:iIp3PUfo9f8PUICapk9cbvNixh3G//E0YcN/H1BS2uU 2 gJOKyEAWqEO3fRvwo7zibw UDP 0.490 74.88.32.5 55690

a=candidate:SwYcgzmLKxjngz0ly88fx3WyBfeLJyFpUSooOFbqiCw 1 AVJa9tLvi/OL7UAMrXtiQg TCP 0.250 76.34.37.65 64896

a=candidate:SwYcgzmLKxjngz0ly88fx3WyBfeLJyFpUSooOFbqiCw 2 AVJa9tLvi/OL7UAMrXtiQg TCP 0.250 76.34.37.65 64896

a=candidate:Kg7LyhMXrFauBM795AMdGHe12oSjdFMnYjThzovUlLY 1 T0uIkUWv0qOZXsAjK9NgDg UDP 0.550 76.34.37.65 4864

a=candidate:Kg7LyhMXrFauBM795AMdGHe12oSjdFMnYjThzovUlLY 2 T0uIkUWv0qOZXsAjK9NgDg UDP 0.550 76.34.37.65 27520

They are called A=Candidates notice there are 10 listed. these sometimes may be referred to as Candidate pairs. notice there are TCP and UDP pairs. once this is passed back to USER B then stun negation begins to take place. and ultimately different patterns are tried until it is finally determined which path is best  to take. in this case User B was able to talk directly to USER A via UDP. So the sniff ultimately showed 192.168.33.74 talking directly to 70.209.48.141.

Because there are so many variables it is hard to say what the voice path is for each and every call. So if someone asks "What is the RTP traffic flow in OCS" the only answer that is possible is it depends. From what I understand so far I believe it is possible for voice traffic to be peer to peer in a multitude of situations. I will start a trace on a 3 person conference call via MOC next. but that may be a day or 2.

Thanks. hope this helps someone.

So you want to set up OCS to work remotely. you read the documentation and try to decide how you want to deploy the edge servers so you can have remote access. There appears to be  a lot of confusion about what is needed and how it all works. so let me see if I can explain a little better.

in a small deployment that wants voice video, live meeting, IM presence to have remote access.  here is what you need and what it does.

• consolidated edge (this has the Access Edge, Web Conferencing Edge, and A. V edge installed on the same box)
• ISA Server for Address book download, Group Expansion while remote, and Live meeting Content Download.

Access Edge, Web conferencing edge, AV edge, all consolidated on the same box. So you can have all 3 roles loaded on the same box but that brings some challenges that we will discuss shortly.

So what does each server and server role do at the edge. Now remember I am not trying to be extremely technical and my description may be a little off but the general traffic flow is correct and the general use of  each server is correct

Access edge proxy's Call Control traffic back to the OCS Front end servers. it also proxies IM and presence traffic from remote users. So When a user is remote or federated all im and presence.nse traffic proxy's through the access edge no other server is involved in the DMZ.

Web conferencing Edge server is a proxy for live meeting traffic to the conferencing focus which is on the OCS front end server in most cases. However the initial setup of the live meeting session is initiated through the access edge. during the initial setup of the conference information is passed about the web URI for the live meeting server to the client so it can find the web conferencing edge server. during this same session initiation the web url to the ISA server for content download is passed to the client as well.

AV edge role is used for Voice and Video proxy so that a user that calls a remote user from inside will initiate the call. which will take advantage of the access edge to establish call control. from this, call control information is passed that tells both clients (the one inside the LAN and the one remote) to send all audio and video traffic through the AV edge. So what this means is all clients in the LAN need to be able to send traffic to the AV edge.  an interesting note is that when a call is initiated internally then the call is peer to peer for the voice and video. When it is remote the video and voice must hit the AV edge. Also when a conference call is established a different path is followed. in a conference call all Audio, video is sent to the front end server so it can be mixed and sent back out.

ISA server

used to proxy the location of he Address book download files, Group expansion files, and finally live meeting content. So will live meeting work with out this in place and will MOC work with out it. Yes it will however for MOC you will not recieve new address book downloads while being remote. And Group expansion may have some problems but the client still works.

As far as live meeting this depends on what functionality you need. this is where any uploaded files would be located so if you want to upload PowerPoints for presentations then this is needed. also if you want to post handouts then it is also a needed resource. Some people suggest just opening up port 443 direct to the front end server but this is not supported not recommended and VERY Risky.

The diagram below shows the three roles on separate servers for simplicity I have left the ISA server out for the time being this will be discussed later. and this is only showing a remote user to internal user. I know my picture is not great. but hey I never claimed to be an artist or even a master at Visio.

So you finally have live meeting up and running and you start to tell everyone it is working and available. The boss takes you up on using it for a meeting and comes back saying this sucks. Now you begin to wonder if it is all it is cracked up to be. What do you do?

Since I work with live meeting all the time there are some things I have noticed that may make conducting a live meeting more successful.

Live meeting is not the same as having all the users in the same room.Nor is a confernce phone the same as having everyone in the same room. So why do people continue to treat it the same. this is just not possible have you ever been on a conference call when you are the remote caller and there are 10 other people in a meeting room with one of those cool polycom phones. And multiple conversations start up in the room YOU HAVE NO IDEA HOW ANNOYING THIS is it is a complete waste of time. you can not tell who is talking and what is important. Live meeting is no different. so here are some ground rules to conducting meetings using technology this is not just about live meeting but the examples will use live meeting.

1. Set ground rules
1. only one person talks at a time
2. No pen tapping
3. No dragging your computer across the table (this is an ear killer)
4. No typing (set your computer in your lap and tap the keys lightly
2. Set 2 people up as the presenter. In case the real presenter is remote and loses connectivity. the other can quickly take over
3. Set a person up to help the remote users. have them watch the icons to see if status changes.
4. pre stage documentation
5. Setup the live meeting before hand have handouts ready make sure presentation is uploaded. Shareing your desktop powerpoint is a process killer.
6. Play with it before the first meeting
7. If video does not work well move on.
8. Mute everyone but he speaker initially
9. if you are a remote user mute your mic before you join.

These are just some of the guidelines I like to use. I will post more as time permits. 

All I have been fighting a battle with vista that has been killing me and many others. There are tons of posts about this issue but i think I have found it and it is an age long problem that has been around for years but we often forget about it.

How many of you have had the infamous I am typeing and it takes a while for the screen to catch up. or outlook hanging and saying "not responding" it seems now that I have found my problems. I was at a point that vista was almost unusable but I kept digging at it. and here is what I found.

First Turn off Aero it  seems to improve things considerably. I did find going to classic view helped a ton. But I still had problems with outlook and other things.

Second I went into Anti-virus and excluded scans on ost's, pst's, and I excluded the outlook.exe  executable.

I restarted my machine and what a surprise I have been functioning back at windows XP levels of speed again it is awesome. 

Microsoft has released a tool to assist in the design and layout of your OCS implementation.  It will create topology designs and give recommended hardware requirements for your deployment.

Planning Tool for Office Communications Server 2007

3/7/2008

Exchange UM with a PBX

When setting up exchange um some things need to be considered. on the initial setup page for the dial plan with in exchange there are sevral options. And if you are not careful and select the wrong option UM will not work. it is important to understand the options. In the URI type you have a choice for 

telephone extension

Telephone extension is what you want to use when connecting to a PBX directly or a sip gateway directly. this may not always be the case but if you know what it is looking for then it will help you decide. for telephone extension it is simply looking to match the extension numbers. and that is all.

E.164 

E.164 looks to see an e.164 number coming from the pbx or gateway. Which can be used in conjunction with a gateway that normalizes to e164 or in locations that already pass e.164 to the gateway.

Sip URI 

Sip URI looks to match to a sip URI. (this is usually used when setting up OCS to exchange UM)

So to begin with when you create a dial plan you have the ability to select the URI type. In most cases if this is direct to the pbx or to a gateway it will be Telephone extension. If you select sip URI or tel URI it is possible for you to have some problems because of the way information is passed from the pbx.

i.e. 6554@10.1.1.1 which is how the pbx may send the information in the sip invite. with 10.1.1.1 being the address of the PBX.  if you select sip uri or sip tel uri exchange will not be able to find this extension or person because the exchange sip uri's will not have the IP of the PBX in them (or at least they should not)

your sip  uri should be your primary email address for best practice (even though some do not follow this) and your tel uri should be an e.164 number i.e. +19139559555 or +19139559555;ext=6554. and since neither will match the previous address with the PBX IP in it. Then the exchange server will return SIP/2.0 302 Moved Temporarily.

So for pbx to exchange you should select extension. so where does the secure verses non-secure come in.

your next selection allows you 3 options unsecured, Sip Secure and Secured. in most cases when you are configuring exchange UM to a pbx or gateway and not through OCS this will be set to unsecured.

Realize that when you set it to secured or Sip Secured you CANNOT HAVE THIS CO-LOCATED on the cas role or hub transport.

Call forking and Simultaneous ring are pretty much synonymous.

in a previous post I talked about call forking. So this week some presentations I went to started talking about simultaneous ring. Whet  I discussed this with the presenters I found that sometimes these 2 terms are synonymous. However some times they are not.

from My understanding True Simultaneous ringing is where you have 2 devices with totally different numbers. When someone calls your extension or DID it comes into the PBX and then your desk phone rings and your cell phone rings near simultaneously. the big key to this is that the numbers are different.

for Call forking it generally is referred to in situations where you have a desk phone and another device with the same phone number. i.e. desk phone and Microsoft Office Communicator. Both the desk phone and the MOC client have the same extension. The PBX forks the call to both devices at the same time and when one is picked up the other is released.

Now it all depends on who you talked to as to the definition. I will try to find more detail over the next couple of days.

thanks

After working with the Microsoft Office Communications Server Team we have come to the resolution that the OCSDesigner included with the resource kit does not and will not work. Even though resource kit tools are generally released without support Microsoft went above and beyond to attempt to resolve the issue. I would especially like to thank Tom Laciano, Sr. Program Manager with the Unified Communications Team for all of his assistance. (http://blogs.technet.com/toml/)

Apperently there are some problems with this tool that is available in the resource kit. So far we have not been able to get it working either but we are trying some different configurations over the next couple of days in an attempt to get it working. Once we have a solution we will post on this blog.

I just found that on the OCS PBX and gateway compatibility list that  Cisco now has 2 routers listed the compatability list specifically lists supported IOS. Now I have not tested this yet but will be interested to see how well it works.

This should be a great feature for Cisco and Microsoft. See the link below it has a link to the cisco documentation but it is a little misleading because it does not take you right there. You may have to select "Cisco Voice Gateways" Then look for Microsoft Office Communication Server 2007 version RTM to Cisco IOS Voice Gateway using SIP with T1 ISDN  They also have a guide for E1 as well.

http://technet.microsoft.com/en-us/office/bb735838.aspx

had an interesting question this week from someone that comes from a traditional Telephony back ground about call forking and the need for it. The reason this is interesting is that in the traditional phone world call forking really was not an option that was used (in most cases not available). when you bought a new phone system you did sort of a flash cut. you might have 2 seperate phones on your desk for a couple of weeks but once fully deployed it was a CUT that happened over night. And you were on the new system. However OCS is not a PBX or a complete PBX replacement

Now I am not from Microsoft so this is purely opinion but Call forking does matter for Microsoft today because OCS is not a PBX and there are some functions that work better on a PBX things like call centers and the multi-line capability are all things that a PBX does better today. And really you have to still have a PBX today with OCS. So why would I want to purchase new phones. Why not take advantage of my current investment. So many want to have a call go to both the PBX phone and the OCS client. and in order to do that you either have to do call forking or have seperate extensions for OCS and do simultaneous ring which does work now. However call forking is not quite there yet we have to wait until the PBX vendors support it. Which will hopefully come soon.

As I said in the previous post Microsoft has done of good job of painting the picture of how OCS works with in their planning guides however it can be confusing because they talk about connectivity to the PBX and being able to send calls to the pbx and your ocs client. This is cool and will work eventually when you pbx vendor starts supporting it.
However today Microsoft only supports Standalone OCS. Now this does not mean they want you to replace your pbx because OCS is not supported as a PBX by itself. what this means is that TODAY you can deploy OCS in a method where all calls to ocs users go to OCS, and all calls to pbx users go to PBX users. but not both.

What this means is that if all your calls go to the PBX then at the pbx a route pattern would have to be established that will send all inbound calls for ocs users across the connection to from the pbx to the OCS media gateway. The intersting thing is you can set up OCS to do outbound calling much easier and do not have to worry about the inbound routing.

So basically you would replace all of your phones with OCS compatabile phones for the users on OCS that are enabled for enterprise voice.

Why is this not supported?

Well the biggest piece has to do with call pickup. Lets say a call comes in and it rang both devices i.e. ocs client and Desktop phone that is associated with the PBX. If I pick up the call on my ocs client there has to be some way for the PBX to know the call has been picked up so it will quit ringing the pbx phone. plus there are some other communications type things that need to be done but this gives you the idea of why it does not work yet. from what i understand microsoft is working with many of the PBX vendors to allow this to take place.