The deadline for taking the CVOICE exam is April 14th 2010.

Everyone in the group must finish the CVOICE exam before that deadline or be left behind!

No ifs, ands, or buts!

I just wanted to say that my wife and I saw Avatar last night and it was freakin’ awesome!

Normally I am not impressed with CGI animation, as it is usually poorly done and they try to mask the bad animation by speeding up the action sequences or making the lighting really dark. For examples of bad CGI watch the first Hulk movie. The Hulk looked more like Shrek to me . Or the first Harry Potter movie when he’s flailing around on the broom – all you see is a cartoon-like blur.

But the animation in Avatar was superb. And the storyline was even better. The concept of controlling another body with your mind, just awesome (a little like the Matrix, but better). And the plot was like a futuristic spin on the American Indian story of how Europeans setters stole their land. I loved how at the end you were rooting for the Aliens to defeat the humans!

Well anywhoo, go and see Avatar 3D while its still in the theater!

What are you doing still reading this blog? Go now!

Its seems to require extra time and effort during studying, but if it works then its worth it. I’m going to try it during my CCVP preparation and let you guys know how it works.

We introduced ourselves and our study goals, and agreed to have a common study plan starting with the CCVP exams. The plan is to meet for two hours every Sunday. We will allocate 2 months per CCVP exam, starting with CVOICE.

We also discussed what workbook vendor we will use for our hands on practice. Unfortunately, not too many vendors have CCVP workbooks and the ones that do have CCVP workbooks do not cover the full CCVP blueprint.

If anyone can recommend a good CCVP workbook that has hands-on labs for the entire blueprint, please let me know.

I remember how relieved I was when I was studying for the R+S CCIE and they announced they were taking ISDN off the blueprint. Good riddance I thought to myself. Who likes sifting through Q.921 and Q.931 debugs anyway? And no companies are ordering ISDN circuits anymore (or so I thought). Why order a PRI for data (which only gives you 23 channels for data) when you can order a full T1 and get all 24 channels for data? And with the advent of DSL and cable modems no one would order a 128k BRI connection to a house or small office either.

In addition to its slower speeds, ISDN circuits were more difficult to configure and troubleshoot than regular T1 circuits. I remember one company I worked for received phone bills in the tens of thousands of dollars for ISDNs circuits that kept dialing or would connect and never disconnect. They were either misconfigured or had bugs, who knows…but who needs this headache?

Well…its back!!! Actually, it never really left. It seems ISDN was, and still is, very much alive in the voice world. So if I want to reach expert status in voice, I’ll guess I’ll have to learn ISDN all over again.

Worldwide Standards for PRI:

T1-PRI – Used in North America – 23 B channels and one D channel

Question: Can the same T1 WIC card be used for a regular voice T1 as a PRI T1?

E1-PRI – Used in Europe – 30 B channels, one D channel, one framing channel.

ISDN-PRI – Nonfacility Associated Signaling (NFAS) – Enables a single D channel to control multiple ISDN PRI’s on a chassis.

Question: How many PRI’s can you control using a single D channel with NFAS?

Fractional PRI – A PRI that uses less than 23/31 B channels. Problem is, you can only have one D channel per interface, so you cannot have multiple pri-groups per interface.

Sample T1 config:

Router2(config)#controller T1 1/0
Router2(config-controller)#framing esf <– frame type esf or sf
Router2(config-controller)#linecoding ami <– linecode type ami or b8zs
Router2(config-controller)#clock source line <– clock source network or line
Router2(config-controller)#ds0-group 1 timeslots 1-12 type e&m-wink-start <– configures timeslots 1-12 to be part of ds0-group 1 and the signal type to be e&m wink-start.

Relatively new command:

network-clock-participate – allows the router to use the clock from the line via the specified slot, WIC, or AIM and synchronize the onboard clock to the same reference.

network-clock-select – allows you prioritize which clock source to use among the clock signals recieved on various ports.

QSIG – Used for transferring signaling information between PBX’s using an IP network.

Sample QSIG Config:

Router(config)#controller T1 0/1
Router(config-controller)#pri-group 1 timeslots 1-24 <– creates a pri-group 1 consisting of timeslots 1-24.
Router(config)#interface serial 0/1:23
Router(config-if)#isdn switch-type primary-qsig <– enables qsig
Router(config-if)#isdn protocol-emulate network <– configures the router to simulate the signaling normally provided by a telco ISDN switch. Used to trick the PBX into thinking its connected to the telco network.

“Voice gateways bridge the gap between the VOIP world and the traditional telephony world (for example, a private branch exchange [PBX], the public switched telephone network [PSTN], or an analog phone).”

PLAR – Private Line Automatic Ringdown – think of the bat phone!

E&M – E&M Ports are typically used to connect PBX’s to each other.

“Like a serial port, and E&M interface has a DTE/DCE type of reference. In the telecommunications world, the trunking side is similar to the DCE and is usually associated with CO functionality. The router acts as this side of the interface. The other side is referred to as the signaling side, like a DTE, and is usually a device such as a PBX.”

Question: How many calls can traverse an E&M port at the same time? Only one?

Understanding Call Legs

At first glance I thought call legs are like router hops, where each hop counts as a call leg, but this is not so!

If you look at the diagram on Pg 164, Figure 3-23 you actually see the Packet Network consists of two call legs (call leg 2 and 3) even though it is only one router hop away. That is because the number of call legs is based the number of router interfaces (or legs) each router has into the voice network.

Would it be safe to say that each interface that has a dial peer configuration on it is considered a call leg?

I’m curious to see how this diagram would look with PBX’s mixed into it. Are PBX’s considered call legs as well?

Interesting note: – Page 180 – Summary “An end-to-end voice call consists of four call legs.” – Is that true for all calls, no matter how many pieces of voice equipment it goes through?

Configuring POTS Dial Peers

Sample POTS Dial Peer config:

Router1#config t
Router1(config)#dial-peer voice 1 pots <– The number 1 is an arbitrary tag. The word pots specifies a pots dial peer as opposed to another type, such as voip.
Router1(config-dialpeer)#destination-pattern 7777 <– This is actually the phone number assigned to this dial peer. Can also be thought of as the destination.
Router1(config-dialpeer)#port 1/0/0 <– This is the voice port the dial peer is reachable out of.
Router1(config-dialpeer)#end

Good to know:

“The tag is a number that is locally significant to the router. Although the tag does not need to match the phone number specified by the destination-pattern command, many administrators recommend configuring a tag that does match a dial-peer’s phone number to help make the configuration more intuitive.”

Poll for the audience – How do you tag your dial peers?

AHA Moment:

You can assign any number of phone numbers to a phone port! The phone attached has no idea what number it is!

Configuring VOIP Dial Peers

Sample VOIP Dial Peer config:

Router1#config t
Router1(config)#dial-peer voice 2 voip <– The number 2 is an arbitrary tag. The word voip specifies a voip dial peer as opposed to another type, such as pots.
Router1(config-dialpeer)#destination-pattern 8888 <– This is actually the phone number assigned to this dial peer. Can also be thought of as the destination.
Router1(config-dialpeer)#session target ipv4:10.18.0.1 <– This is the ip address of the destination router, CUCM, or gatekeeper which this phone number is routed to. If the IP is that of the CUCM or gatekeeper, the CUCM or gatekeeper will return the IP address of the remote terminating device.
Router1(config-dialpeer)#end

It is recommended to point the session target to the loopback address of the destination router, as opposed to a physical interface address which may go down.

Matching Inbound Dial Peers

According to the CVOICE book, there are three elements in the call setup message that are matched against four configurable dial-peer command attributed:

1. Called number dialed number identification service – this is the call-destination dial string, and it is derived from the ISDN setup message or channel associated signaling (CAS) DNIS.
2. Calling number automatic number identification – This is a number string that represents the origin, and it is derived from the ISDN setup message or CAS ANI. The ANI is also referred to as the calling line ID (CLID).
3. Voice Port – This represents the POTS physical voice port.

Question: Is DNIS and ANI information available from a plain analog POTS connection, or only on digital ISDN or T1 connections? If not available from POTS connections, how is this information determined?

In honor of Superbowl Sunday, we’re kicking-off our CCVP/CCIE Voice Mastermind study group this Sunday 02/07/10 at 3pm.

Are you studying for the CCVP or CCIE Voice? Why go it alone?

You can learn much faster, have more fun, and gain insights that would take you 10 years to learn on the job.

We’re looking for like-minded individuals who are self-disciplined and serious about earning these certifications. Lazy people who make excuses and give up at the 1 yard line need not reply.

Check out the group requirements here:
http://strategicinfra.com/blogs/ccna-v_ccvp_ccie-v_voice/?p=99

If you’re interested, email your skype ID to ccievoice@strategicinfra.com

Thanks,
Gregory Gombas
CCIE#19649 (R+S)

There are three main methods used to determine voice signal quality:

• MOS - Mean Opinion Score – Has listeners rate pre-recorded conversions on a scale from 1 to 5 (5 being best).

• PSQM - “PSQM Software usually resides with IP call management systems, which are sometimes integrated into SNMP systems”…”Measurement is made by comparing  the original transmitted speech to the resulting speech at the far end of the transmission channel.” It seems the benefit of of PSQM is that measurements can be made in real time, but it is only 90% as accurate as MOS and “does not take into account the jitter or delay problems experienced in packet-switched voice systems”.

I’m confused as to how it does not take into account jitter and delay problems? I mean if it compares voice quality at both ends of the channel, wouldn’t it detect if parts of were clipped?

• PESQ – Similar to PSQM in that it compares the voice signal at both ends of the channel. It “combines the best of the PSQM method with a method called PAMS.” Unlike PSQM, however, PESQ can take into account jitter and packet loss. “PESQ is the preferred method of testing voice quality in an IP network.”

What I would like to know is how PSQM and PESQ is used in the real world? Does Cisco integrate these measurements in their CUCM system, or are they separate appliances?

Transporting Modulated Data over IP Networks

This section covers sending Fax and Modem over VOIP networks. While reading this section I kept thinking to myself how funny it was that the whole purpose of a modem was to take a digital signal and convert it to an analog signal so it can be sent over a PSTN network. So now when we move to a VOIP network, we have to convert the signal into digital again! So now we’re taking a digital signal, converting to analog, converting to digital, back to analog, back to digital again!

I wish I could comment more on Fax over VOIP, but I fell asleep half way through that section. I will have to use toothpicks to prop up my eyelids and reread it.

DSP

“DSP is a specialized microprocessor designed specifically for digital signal processing.”

If I understand correctly a DSP is almost like an ASIC in that its a specialized processor designed to take the burden off the main CPU of the router. It performs such functions as:

• Transcoding – Converts from one voice codec to another.
• Voice Termination – Translates the voice signal from the TDM network to the VOIP network
• MTP – Seems to terminate two voice streams using the same codec but different packetization schemes. I’m not too sure what this means? I need to find some examples of different packetization schemes.
• Audio Conferencing – Acts as a conference bridge – pretty self explanatory.

What I would like to know is which of these DSP functions can be done on a server instead of the router? And why would you do these functions on a router, which has limited memory and processing power, when you can do them on a server that has seemingly limitless amounts of memory and processing power?

Configuring Conferencing and Transcoding on Voice Gateways

One of the steps involved in configuring transcoding and conferincing on a voice gateway is to enable SCCP.

“The Cisco IOS router containing DSP resources communicates with CUCM using SCCP.”

I thought SCCP was only used for Cisco IP phones to communicate with the CUCM server?

Here is a sample SCCP config with my comments:

Router1(config)#sccp local fa0/1 <– sets the source IP of SCCP packets.

Router1(config)#sccp ccm 10.1.1.201 identifier 1 priority 1 version 4.1 <– adds CUCM server with IP 10.1.1.201 to list of available servers .

Router1(config)#sccp <– enables sccp on the router globally

Router1(config)#sccp ccm group 1 <– defines the sccp group

Router1(config-sccp-ccm)#bind interface fa0/1 <– What does this command do? Why does the group need to be bound to an interface?

Router1(config-sccp-ccm)#associate ccm 1 priority 1 <– Associates the previously defined ccm server to the group. But what does this priority command do? And how does this compare to the priority assigned to the sccp ccm command above?

Router1(config-sccp-ccm)#associate profile 1 <– Associates a previously defined dspfarm profile to the group.

Router1(config-sccp-ccm)#register XCODERouter1 <– Registers with the CUCM server using the name XCODERouter1.

Business Case for VOIP – What no mention of the cost savings on MACs (moves, adds, changes)? Or how about the cost savings of less cabling infrastructure? I remember these being the biggest selling points when IPT first came out. Maybe they’re just so well known its almost assumed? Or could it be because you really don’t get much cost savings from these?

Multipoint Control Unit (MCU) – I’ve never actually seen one of these in the real world. I need to find some examples of MCU’s.

Voice Signaling Protocols – Good lord there are so many signaling protocols!

A buddy of mine had this to say:

“Thats what I love about standards, there are so many to choose from!”

What I would love to see is a drawing of a sample network where all these protocols are used, showing where they are used, and how. For example, you can show a CUCM server with arrow pointing to Cisco IP phone that says SCCP communiction, and maybe an arrow pointing to a router showing MGCP communication. Maybe if I get bored one day I will draw such a diagram myself…

What I need to do is get a better understanding of which protocol to use in which situation.

H.323 – You just gotta love protocols within protocols! I’m curious how H.323 is used these days? The last time I saw it used was with Microsoft conferencing software…

MGCP – Any relationship to SCCP?

SIP – I can’t wait to deep dive into SIP. It seems like its all the rage these days. From what I understand non-Cisco IP phones typically run SIP, so what I would like to see is how a voice network looks like when you are running some phone SCCP and some phones SIP?

Most notable sentence: “The primary goal for all four of the previously mentioned VOIP signaling protocols is the same – to create a bidirectional Real-time Transport Protocol (RTP) stream between VOIP endpoints involved in a conversation.”

I find it interesting that the actual voice traffic does not necessarily need to pass through the routers that are involved in the call setup.

In my mind (which is entrenched in the routing and switching world), I guess this is kinda like how BGP speaker can advertise a route with a next hop other than itself, so the traffic doesn’t actually have to pass through itself to reach the destination.

Here is an interesting post about how IPT is a dead end field:

http://bit.ly/4yYkVM

Man I hope he’s wrong!