Call Manager 9.1 - Inbound vs. Outbound routing

[dudecrush]

All - I'm new to Cisco admin (7 months), after spending 9 years as an Avaya Definity systems admin. Naturally, these two systems work in fundamentally different ways. I find myself unlearning my Avaya and relearning new Cisco quite often. I make mistakes because when I have a problem, I fall back to Avaya. What has saved my bacon as a Cisco admin is some previous experience (different employer) and decent documentation left behind by my predecessor. Following the idiot notes has one shortcoming - I don't know WHY it works the way it does - I just know that it does. Like most admins, I actually want to understand what I'm doing. So...

I'm trying to get my head around how inbound / outbound call routing works. I "kinda" get it, but when I try and apply it I just get spun around. Cisco's terminology doesn't help - given that names like Location and Region are always throwing me - not the least of which because the terms seem interchangeable.

a) How are Partitions and Calling Search spaces related? Is one used for inbound and one for outbound, or is it a bit of both?

b) How does an inbound call "hit" a phone? I think its PRI > Voice Gateway > Call Manager > Translation Pattern / DN > Device. Is that correct? Am I missing any steps?

c) How does an outbound call make it "out"? I think it's Route Filter > Route Pattern > Partition > Route Group > Route List > Gateway. Is this correct, or did I miss something?

d) Why do I have partitions, when none of my phones seem to use them? All the directory numbers have the Partition set to "None". IF Partitions are a collection of route patterns, you would need that set. But...only the calling search space is set on the DN. To me, the phone shouldn't work because it can't reach the route patterns. I don't get it.

e) What the heck is the difference between a location and a region? Both seem to control bandwidth, so where would you use one over the other? I'm guessing they're used in conjunction?

f) How can I "trace" an outbound / inbound call from cradle-to-grave? In Avaya-land, I could open up the terminal emulator and type "List trace station xxxx" and watch a call go from hop-to-hop. This made it easy to know where the call died. Is this a function of RTMT, or can I run that in Call Manager some place? (I've learned to run call traces on routers, but it throws so much crap it's really hard to parse.)

g) Call Manager's CDR reporting is so-so - but what I really need is Call Accounting. Is that in Call Manager, or should I used RTMT for that?

h) I want to use RTMT like the Definity GEDI, but it's confounding me. I'm guessing that RTMT is not meant to be used in the same way as GEDI?

Any help or hard-won knowledge is appreciated. And yes....I will be getting some training next year, but I want to learn all I can before then.

 

[gnrslash4life]

a) How are Partitions and Calling Search spaces related? Is one used for inbound and one for outbound, or is it a bit of both? - Numbers reside in the partitions. CSS's have access to the partitions you let it. Used for both inbound and outbound as well as internal.

b) How does an inbound call "hit" a phone? I think its PRI > Voice Gateway > Call Manager > Translation Pattern / DN > Device. Is that correct? Am I missing any steps? - That is one way, yes. Depends on what protocol you are using (MGCP, H323, or SIP).

c) How does an outbound call make it "out"? I think it's Route Filter > Route Pattern > Partition > Route Group > Route List > Gateway. Is this correct, or did I miss something? - Ive never used filters myself, but it would start with the CSS, then partition, then Route Pattern, etc.

d) Why do I have partitions, when none of my phones seem to use them? All the directory numbers have the Partition set to "None". IF Partitions are a collection of route patterns, you would need that set. But...only the calling search space is set on the DN. To me, the phone shouldn't work because it can't reach the route patterns. I don't get it. - None is essentially a default partition. If the DN can reach the partition of the Route Pattern then it will work.

e) What the heck is the difference between a location and a region? Both seem to control bandwidth, so where would you use one over the other? I'm guessing they're used in conjunction? - Location = Bandwidth, Region - Codec usage.

f) How can I "trace" an outbound / inbound call from cradle-to-grave? In Avaya-land, I could open up the terminal emulator and type "List trace station xxxx" and watch a call go from hop-to-hop. This made it easy to know where the call died. Is this a function of RTMT, or can I run that in Call Manager some place? (I've learned to run call traces on routers, but it throws so much crap it's really hard to parse.) - CUCM Serviceability, Dialed Number Analyzer.

g) Call Manager's CDR reporting is so-so - but what I really need is Call Accounting. Is that in Call Manager, or should I used RTMT for that? - Call Accounting needs to come from a 3rd party.

h) I want to use RTMT like the Definity GEDI, but it's confounding me. I'm guessing that RTMT is not meant to be used in the same way as GEDI? - Not an Avaya person so I cant say if its the same but it provides 2 things. 1)System Health 2) Troubleshooting tools. Big time on collecting trace files.

Any help or hard-won knowledge is appreciated. And yes....I will be getting some training next year, but I want to learn all I can before then.

Certifications:
A+
Network+
CCENT
CCNA Voice
TVOICE
CAPPS

 

[dudecrush]

gnrslash4life - This is a follow-up to your post from Sept. 15. I've learned a few things and answered some of my own questions since that post, and I have 2 follow-up questions:

Question 1:
In b) above, I asked how and inbound call "hits" a phone. What I've learned is: Gateway > Calling Search Space > Partition > DN. Since all of my DN's are in the <NULL> partition, all of my calls are reachable by anyone. Bad design.

Anyway, you mentioned that the way calls route depend on the protocol: H.323, SIP or MGCP. That's interesting. Since I (mostly) use MGCP, what I described above then should be the way MGCP routes a call. Can you explain the path H.323 and SIP takes route an inbound call?


Question 2:
In c) above, I was questioning how a call makes it out. This one I'm still unsure on. I was told that the outbound process is this:
> The CSS of the device and the CSS of the DN are checked (which one of these takes precedence?)
> When digits are dialed, it's checked against the CSS
> The CSS checks the order of the partitions in the dependency records
> Route Pattern > Route List > Route Group
> Route Group > Voice-Gateway

> Is this the correct path?
> Is outbound routing dependent on the protocol like inbound routing?
> Would the Dialed Number Analyzer give me the correct order of outbound routing?

 

[gnrslash4life]

All protocols route inbound calls from a CM perspective essentially. H323 and SIP bring in things like dial peers and you have more control over things. MGCP to me is easier to deal with but ive met many people who disagree. I started out on MGCP and only work on SIP now. To each their own.

You are pretty much correct on your outbound thinking. Translation patterns play a roll too. Also the CSS on the DN takes precedence over the device CSS.

DNA is a great tool and yes, it will show you all available routes.

Certifications:
A+
Network+
CCENT
CCNA Voice
TVOICE
CAPPS

 

[dudecrush]

Hark! - You know SIP!

We are moving all of our 40+ MGCP and H.323 locations to AT&T SIP, and it's putting a pit in my stomach. I feel like I'm getting dropped into a meat-grinder with this project, and I'm not getting much help from my management. What "gotchas" did you encounter in your conversion?

 

[gnrslash4life]

I actually work with AT&T SIP quite a bit too. They arent bad. Nothing funky to configure. Some of the smaller telco's are a pain with SIP fyi. If you can handle the H323 config of the router, you can handle SIP. Its very similar.

Certifications:
A+
Network+
CCENT
CCNA Voice
TVOICE
CAPPS

 

[dudecrush]

Awesome! A question...

When we tested, AT&T wrote a translation rule for their test numbers. When you ported numbers to your SIP trunk, did you need to write additional translation rules on your SIP gateway for your DID blocks, or was that not necessary? We have multiple non-contiguous blocks, and it would be a real pain to try and write them all.

What else do I need to consider on the routers, besides dial-peers

What about remote gateways? What else should be considered?

 

[gnrslash4life]

No translation rules. I always had them send me 10 digits. But it depends on how your system is designed. Its hard to say if there is anything in particular to look for. In my experience though, use SIP gateways, not H323. When I worked in a NOC I saw so many clients running SIP into a H323 and it just baffled me. Didnt usually cause issues but ran into some weird ones.

Certifications:
A+
Network+
CCENT
CCNA Voice
TVOICE
CAPPS

 

[dudecrush]

OK - It's the translation rules that I was trying to understand. I have 40+ locations, and (mostly) we have a 3-digit translation pattern prefix attached to each location's 4-digit extension. I wasn't sure if I could just "leave things alone" and let AT&T send me 10 digits, or if I needed to break that down using translation rules on our SIP gateways.

What about dial-peers? Did you need to do anything with those? We set up a dial-peer for the test numbers, but I'm unsure if I need one for each of out branch locations.

 

[gnrslash4life]

You could do the translation on the router, or do it in CM.

Dial-peers depends on what you got going on. You could get away with typical dial peers for incoming/outgoing calls and let CM handle it as well. But there is no definitive way to do it since every company is different.

Certifications:
A+
Network+
CCENT
CCNA Voice
TVOICE
CAPPS