CCIE Voice Home Lab Gear

I personally prefer to have my own lab at home rather than use any rent racks so that I can turn it on in any moment and try anything I want to. I've spent a few months and quite a lot of money to obtain all necessary equipment, but now I have almost completely authentic lab where I can study and train just everything from the Blueprint. Below is my gear:

HQ Site:

CISCO 2811

WIC-1T for FR

VWIC-2MFT-T1 for ISDN

PVDM2-16

WS-C3750-48PS-S

Branch1 Site:

CISCO 2811

WIC-1T for FR

VWIC-2MFT-T1 for ISDN

PVDM2-16

Branch2 Site:

CISCO 2811

WIC-1T for FR

VWIC-2MFT-E1 for ISDN

PVDM2-8 x2 - we need two dsps at least for one of the sites so that we can play with the Conference Media Resources. It can be PVDM2-32 for example since this module has two dsps on it. PVDM2 dsps use special firmware for conferencing and that is why it can't share dsp with any other function (transcoding, tdm-voip translation and so on). 

AIM-CUE

PSTN-FR Switch:

CISCO 2811

WIC-2T

WIC-1T

VWIC-2MFT-T1

VWIC-2MFT-E1

PVDM2-16

Phones:

7960 as a PSTN phone

Mix of 7941/7942/7961 phones for the sites

I use one 3550 PoE switch instead of both HWIC-4ESW (Branch1 and Branch2) because it's much much less expensive and there are only a few small differences in ports configuration between them. Also I use first genereation VWIC cards rather than VWIC2 for the same reason (we only need to remember one command for VWIC2 - card type e1/t1). And for sure I could use spare ports from the VWIC-2MFT-T1 and E1 for the Frame Relay and it's actually a bit easier - you don't need to purchase WIC-1T cards and cables, but I have had a few such cards and decided to use those. Also I have only one 7961 phone and I rarely have to use CIPC for features testing, because those demand more than two phone buttons.

And by the way, just a small hint. I was trying to get rid of this loud annoying noice from the 2811 routers (switches are pretty quiet) using different ways, but without any success - they didn't cool enough. And all my efforts end up  with this guy - Scythe Kaze Q 5.25'' - 12-channel fan speed controller. I've just removed Network Modules panels and plugged all the fans to the controller. So now I can control fan's speed in the range of 5V - 12V and it's almost always enough to use a minimum (5V).

CUCM Based Presence

In order to make corporate directory show up user's presence status telephone number attribute in the end user's settings has to be exactly the same as the one on the line you are going to monitor. For example, my CUCM is integrated with AD and Hugo Reyes has 1002 extension in the telephone number field and Jack Shepherd has full E.164 number. All the lines are configured with 4-digit extensions. Once Presence Groups along with Subscribe CSS have been configured I can only monitor Hugo (small phone picture instead of keypad). Having numbers different from the actual extension work fine though for Missed/Received/Placed calls because there are extensions you've gotten calls from in the way they configured in the line settings in there.

BLF Speed Dials only deal with SUBSCRIBE CSS. If the Subscribe CSS does contain necessary partition you'll be allowed to monitor the line.

BLF Call Lists deal with combination of SUBSCRIBE CSS and Presence Groups. In order to monitor someone in the call list/corporate directory your Subscribe CSS has to have watchee's line partition and also your device Presence Group should be allowed to monitor the watchee's line's Presence Group. You won't be able to monitor call list if you have appropriate Subscribe CSS but your PG is disallowed to monitor necessary PG and vice versa.

NX-OS cache

Not related to CCIE and Voice at all, but just quick note to remember.

I've decided to add a few new Nexuses 5548 (completely new system for me) into the ACS the other day and faced with weird for me issue. I've performed the usuall procedure: create new device in ACS server using our standrad shared key and copy standard piece of code from some router to the Nexus. Once I did it and tried to login I've got incorrect login or password message. I checked everything a few times without any luck, then I opened ACS tacacs log and saw invalid shared secret message. Ok, but how is it even possible, I copied it from another router. I started to check another Nexus configuration, which has already been added to ACS and noticed that the encrypted shared key in configuration looks completely different from the one I added from router. Ah, that was the issue - IOS and Nex-OS have completely different caching algorithms, that is why it's not possible to copy encrypted keys from one OS to another. So after I've typed the raw key in I was able to log into the Nexus successfuly using my AD account.

Unity Connection 7.0.1 Temp Fail

A quick note about the home lab gear.
My home PC has a powerful AMD CPU inside, which I've purchased specifically for the Lab Virtual Machines but that seems to be a big mistake for me now. I have been getting a lot of issues with my virtual servers running in the machine with AMD CPU.
First of all I couldn't even install any Linux based servers: CUCM, Unity Connection and CUPS, so I had to install those on my Intel based Laptop and move to the PC, but since my Laptop is quite old this process has taken almost whole day from me.
Next I've faced with this IP Voice Media Streaming service issue: it just fails a few seconds after activation, so I couldn't test any Software Media Resources and Annunciator and MoH for example don't even have any hardware option like all the other resources!
Ok, I have found out the workaround in Internet after a few hours of googling:

• Download any Linux image (I have used the CentOS one)
• Inser it into a virtual server's CDROM
• Boot the server from the CD
• Choose recovery process among the options
• Navigate to the /lib/modules/cm_ipvms
• Duplicate all the folders with .i686 word in the name but use .athlon word instead
• Copy all the contents from .i686 folders to the newly created .athlon folders

The next issue is Unity Connection related, in fact it just doesn't work! I can integrate it with CUCM/CUCME, can create/import users, can even leave a message, but whenever I try to login Connection Conversation Manager just crashes! I enter the PIN and see Temp Fail at the phone screen. After a few minutes the service gets up, but crashes again if you try to enter PIN again.
I've decided to move this VM Machine back to the Laptop, added it to the Shared folder and now I can manage it from my PC though it actually runs in the Laptop, that did the trick: Unity works fine now and I still have centralized servers management point.
As I know all these issues have been resolved in 7.1 versions, but I want to train with the exact Lab versions which are 7.0.1. So if someone would like to build its own home lab for the CCIE Voice lab preparation, you should only use Intel processors for your VMWare workstation/server!

CME 7.0.1 Default Software Conference Behavior

What is the default behavior of regular software ad-hoc conference in CME system, before we'd started to play with keep-conference command in the ephone or voice register pool modes?

SCCP:

Last participant is being removed once the conference softkey pressed
EndCall softkey just disconnects all the participants no matter if there are local participants left in the conference or not
Going on-hook works in the same way as EndCall does

SIP:

Conference remains active in any way

If there is Conference Hardware command issued in telephony-service mode sccp phones lose the ability of mixing streams localy and show up Cannot Complete Conference if you try to add a third participant in the call while there is no Conference Bridge configured.

SIP phones don't support MeetMe conferences. For sure they can participate, but can't organize one. They don't even have MeetMe Softkey.

CME 7.0.1 Directory Service

Directories as well as Services require the ip http server to be enabled and if it's not a Host not found message appears on the screen once Directories or Services buttons pressed.

Quite useful directory related service dnis dir-lookup command is only has influence on called number, not the calling one, which actually confuses a bit. Why do we need it?

For example you are the call center serving plenty of different customers and each customer knows one of your numbers so in order to let the operator know what customer calling him right now we can use below configuraration:

telephony-service

     directory entry 1 555777111 name Customer 1

     directory entry 2 555777222 name Customer 2

     directory entry 1 555777333 name Customer 3

     service dnis dir-lookup

And if a customer dials 555777111 operator will see calling number on the screen as usual, but he also will see Customer 1 message at the bottom, so that he can prepare himself and answer in the proper way.

Frame-Relay traffic-shaping

In order to enable frame-relay traffic shaping issue the frame-relay traffic-shaping command in the main interface mode. When the command frame relay traffic-shaping is applied on the main Serial interface, a separate queue is created per Virtual Circuit. Note that once enabled, a CIR of 56Kbps and a Tc of 125ms applies to all active PVCs.