logs archiveBotHelp.net / Freenode / #2600hz / 2015 / August / 7 / 1
fseesink
Man how I wish that Kazoo would support the Cisco 8800 series of SIP phones like the 8851 and 8861.
Had an interesting moment yesterday. Been part of a project involving migration to Verizon UCCaaS (*) for the better part of 6 months now (no joke). [(*) Not my choice, as I strongly advocated against it, but I lost that battle.]
At one point I observed I could have built a solution from scratch by now, to which I had more than one person ask, So why dont we start doing that? The frustration level has gotten that high.
Of course, Id love to see us stand up a Kazoo cluster, as we have the resources. But the challenge, as many of you may know, is the investment in handsets. Since the decision was made, a lot of $$ was spent on these Cisco 8800 phones, so the idea of having to buy all new again simply wouldnt fly. But if we could re-use them (much like when I had hoped the Cisco 7900 series would be supported for similar reasons), that opens the door.
stormqloud_
Does this UCCaas install the kitchen sink, or just ship it to your dock?
fseesink
But while these new phones are at least SIP vs. SCCP, looks like architecturally they operate in a similar manner. That is, do DHCP on boot, get response including Option 150 for the IP of TFTP server, TFTP to said server and download filename CEPaabbccddeeff.cnf.xml where aabbccddeeff is the MAC address of the phone, and configure.
stormqloud: UCCaaS stands for Unified Communication and Collaboration as a Service, or as I call it, Cisco CallManager in the cloud. Verizon is pimping this as their hosted VoIP platform, with multiple datacenters on east/west coast, and among the many words you will hear them use is multitenant, though as I have written here recently, its a bit like that scene in Princess Bride&You keep using that word. I do not think it mea
what you think it means. But theyve chosen to build their platform on CallManager.
stormqloud_
why do they not work with kazoo if they are sip
I dont; remember why anymore because it was years ago but there was an issue in using some cisoc models in a hosted pbx manner
fseesink
Well, Kazoo doesnt have official support for this series as yet, not in their provisioner, and I cant find anyone whos made it work yet.
stormqloud_
ah, the provisioner is a diffeent matter for sure
stormqloud
the 7960 & 7940, somehting about remote provisioning was always an issue unless the remote provision came from a local cisco router
fseesink
Thing is, if 2600hz/Kazoo officially supported this series of phones, Id have a shot at showing how we have options if people do finally have enough. (Its far too long a tale to tell, but I argued against this solution for a very long time, and unfortunately everything I warned about& and even some things I didnt& have come to pass. Its actually been worse than I expected which is saying something. But while I can offer my
professional opinion/eval, Im not at the level to make the decision, and it was decided to go this route. Once done, I just had to shut up, as at this point it comes across as I told you so, and I dont want to be that guy. But now others are seeing what Ive been warning about, and its amazing how much more quickly theyre becoming frustrated with this whole process.
UCCaaS is a subscription-based model, a little like 2600hzs, where you pay $/mo/device with different tiers (basic, enhanced, zirconium, ultra hot gold& :-) ), from a basic phone in a waiting area to one offering voicemail, mobility, softclient support, etc.
stormqloud
some big institutions would rather do the "nobody ever got fired for buying IBM" strategy. Which is totally out of date as an expression now.
but the concept is the same
fseesink
So if Kazoo supported these phones, and hell truly broke out, we could dump them and move. But you have to be able to move your existing hardware, namely the phones. And in our case, wed likely have to work with 2600hz to see about getting SBCs on-prem to handle 911 failover.
stormqloud
a CXX that knows nothing abotu phones will not get fired for buying Verizon even if it fails
they will get fired if they go with some small company and it fails
fseesink
exactly (buying Big Blue). And in our case, part of the reasoning was that Verizon has a state contract, meaning to go this route avoided having to do bidding (which sucks at state govt level).
stormqloud
y in the US you have the fed govt price thing also
you can;t fight against that. It's jsut tilting at windmills..
fseesink
but part of the reasoning, which I tried hard to explain was a total fallacy, was that it was Cisco, and were a Cisco shop. Well, if you know Cisco routers and switches, that does you spit when it comes to CallManager. The time youll invest learning that you could learn anything else, and in most cases, in less time.
stormqloud
y I love cisco for decades..
but there are certain platorms I will not touch
cisco ASA devices (shoot me first) and callmanager
fseesink
Well, in our case, were an ISP/datacenter with the resources and the skills to host our own platform if need be. So the company failing aspect isnt a solid argument, especially when youre talking open source. And 2600hz has built a model I think is solid. They give the keys to the kingdom if you want it. They can host for you if you dont want to, or theyll help you move on-prem (and back again if later you opt to do that). They
have a setup that should make anyone feel safe.
Gee, thanks. one of the projects top of my list is implementing some ASA 5515X units to replace aging VPN concentrators. But yeah, I know what you mean.
:-)
stormqloud
It was when they switched the syntax of the global command years ago
fseesink
I mean, who doesnt love me some Java client software in 2015, right?
*cough*
stormqloud
the netmask format was changing
everytime I upgraded an ASA it became a nightmare of updating rules
fseesink
Oh, thats nothing. The change to what they call an object oriented configuration is the kicker.
I think it will help in the end, but the transition from the old config format to the new is hell if you have to actually bring over a config.
stormqloud
in all seriousness I have a cisco external guy that I use for any ASA stuff. I decided years ago I didn;t want to even know about the ASA anymore. I was never goign to buy another one.
fseesink
A lot of things are just more complicated than they need to be.
stormqloud
I gg hang out iin a data center a freeze for a bit. Later.
fseesink
Yeah, were not so lucky. Again, state agency in essence (under higher ed to be precise). Were on fixed salaries, but we have to be jack of all trades here if we want to get anything done.
neurosys
fseesink: I made progress with teh Cisco 8861
Just keep distracted lol
fseesink
neurosys: hehe.
yeah? Did you get my message yesterday?
neurosys
Nope I missed
fseesink
Ill help where I can with this. I have over a dozen 8851s sitting on a few desks across from me, which every time I see them makes me think of this: https://www.youtube.com/watch?v=atUUjSLMSiM
neurosys
fseesink: The same issue applies: the phone isn't appreciating the traffic originating from somewehre else
fseesink
Theyre burning in, by which I mean that for some odd reason, 30-50% of them when first booted/configured, will for the first few days go into a registering& state that only gets resolved by resetting the phones, and then it settles out. Of course, youd think Verizon (you know, the PHONE company), would have some solid handle on what is the current latest enterprise series of phones that work with their solution. But their guy
clueless as to why some of these phones do this while others dont. So before we deploy them to replace 7961G-Ges, were letting them sit a few days to see if they behave.
neurosys
Im sure there is a cryptic XML attribute toy adress this. but i havent figured it out just yet
Outbound works fine
fseesink
(copy/paste from yesterday dm) Been meaning to ask you, any more luck with that Cisco phone? When I asked if you wanted a sample of the cnf.xml file that CUCM generates (what the phone pulls down via TFTP), you mentioned you pulled the config from the phone itself. How did you do that? I mean, I know the web UI on the phone, but that doesnt provide you the XML config file as far as I can tell. It tells you what certain things are set to, s
but it doesnt have all the tags/etc. (and there are a LOT of them, at least in CUCM 9.1).
neurosys
fseesink: the phone can be accessed via SSH so i can dump the config and get console debugging...
fseesink
ahhh
neurosys
documentation is scarce since cisco likes to charge you your first unborn child to know how their crap works
fseesink
huh& just tried to my phone and got a rejection.
SSH that is
neurosys
fseesink: SSH login creds are set at the XML
fseesink
neurosys: so you had to first push a config to the phone, then you could ssh?
neurosys
yep
fseesink
ah. Nothing like simple& and Cisco sure makes it nothing like simple. :-)
neurosys
yeah when an incoming call comes into the 8861, it rejects as the Request URL not found.
*URI
It doesnt like that it's not coming back from the realm
fseesink
*sigh* I really wish I could have convinced folks. We couldve avoided all this and been able to buy the right kinds of phones. But just as I was dealing with an on-prem CUCM earlier, now Im stuck with a cloud version and in both cases these #$%^& phones.
I tell ya, the non-Cisco VoIP world could do some damage if they did support these things. I mean, truth is, the suits only care about appearances anyway. So if they have a Cisco phone on their desk, most likely could care less if you had squirrels running in cages on the backend doing the actual work.
Many years ago, when we first got our CUCM, I was still working with Asterisk (pre-FreeSWITCH days), and I got things to the point where I had a TFTP server setup, had the config files formatted, had phones registering, and was able to do basic things. But it was very much under the hood, doing things via CLI, etc. And even now, sounds like nothing has changed much. I mean, youre doing similar things, then SSHing in to get at the
configs/debug logs, etc. But until theres clean UI support where you can provision one of these like you can some of the Yealink or Polycom units in Kazoo (and I have to say, I really like the new Monster UI vs. the Kazoo UI for basic PBX setup), theres not much chance of going in and replacing CUCM backends in places where currently theres maybe CUCM Express running on a router, or CUCM Business Ed., let alone going after large
installations.
Speaking of UIs, I sent an email several days ago to the same guys who expressed frustration yesterday, showing them the Web UI of CUCM 8.0.x vs. UCCaaS 9.1.x (keeping in mind that the 6-mo. tick tock release cycle of Cisco CUCM went 8.0, then 8.1, 8.5, 8.6, 9.0, 9.1, 10.0, 10.5, and the recently announced 11.0), so basically 4 major revs over several years& and you couldnt tell them apart other than the version # listed. Then I sho
screenshots of the Kazoo UI vs. the new Monster UI, and thats just in the past 2 years.
Just shows you how much faster the non-Cisco VoIP world is moving. Dont get me wrong. Im all for leaving a UI alone if its good, as Im no fan of change for changes sake (you listening, MS Office devs?). But CUCMs Web UI blows.
Its overly complicated for smaller institutions. I mean, if youre a huge corp. with dedicated VoIP staff, then by all means look at CUCM with its clustering ability, etc., as standalone Asterisk and FreeSWITCH arent really built for that. But in my mind you really need to have somewhere in excess of 2K endpoints before you have to go that route.
And Kazoo looks to be architected to handle those large installations well, using Kamailio and FreeSWITCH for their respective strengths, among other things, and using Erlang as the glue. (As a language wonk, I appreciate the language choice, as Erlang was built for this kind of thing, though admittedly I have yet to buckle down and wrap my head around that language& functional languages are a different breed.)
Hey, totally different question for the peanut gallery: is anyone integrating their LDAP/Active Directory into their setups, and if so, how? Just as a corporate directory listing type setup? Full on integration where you deifne your users in LDAP/AD and tie it into your PBX?
Its a long story, but a shorter version is that I ask because yet another place where things have fallen down in the UCCaaS setup is the whole notion fo multi-tenancy. They tried playing games in that regard by giving all devices unique, fake DIDs of a pattern (e.g., 101-101-xxxx for site 1, then 101-102-xxxx for site 2, etc.) as well as mapping local extensions to them. Then they setup calling partitions/etc. so phones of pattern1 dial
a local extension only reached other phones within that pattern, etc. But its a cludge at best. And it doesnt account for things like user accounts, voicemail, etc. Then it turned out that while CUCM can sync to mutiple LDAP/AD stores, it can only sync to ONE for authentication purposes. So if youre pimping this product as a multi-tenant solution, youre in for a world of hurt. If you have two clients, each with their own LDA
store, and each has a John Smith, they canot both have the same username. And so on.
Thus far, every issue has been tackled with what I call a nails and duct tape solution. And this is because CUCM was clearly never architected to be truly multi-tenant. If you setup a standalone FreeSWITCH server, having multiple clients with the same local extensions is not an issue, and ditto for Kazoo. Clients are truly separated. But anyway, I digress. I was just curious if anyone is using either the hosted solution or running their
Kazoo cluster, and does anything with directory integration.
bpdavis
does anyone know if there's a way to preserve a from header param like ";user=phone" when a call is bridged by kazoo?
bazyar
hi folks!
Can anyone tell me if they have successfully implemented shared call appearance with Kazoo ?
stormqloud_
you mean "key system" type shared lines?
Line 2 is for you http://www.lighting-gallery.net/gallery/albums/userpics/10519/normal_phone.JPG
Anyone get Freeswitch throwing off RTCP info
bazyar
stormqloud - yes, basically, that old key system type setup.
I know, it seems dumb, but a lot of people don't want to change the way they do stuff!
stormqloud
yrs, I know. It's like when you go see a client and the clients wife is 80 year old receptionist
they are not going to change
bazyar
Right
stormqloud
in canada particularly this is common to have shared line appearances. Norstar/nortel key system which sold all over do that
bazyar
I know polycom and grandstream support it but I'm not sure of details around how you'd set this up in Kazoo
stormqloud
not sure either. can you fake it with BLF somehow
bazyar
I tried to get them to switch to using 1-button park/unpark but they don't like it.
stormqloud
shared lines isn;t only in the phnes tho, requires server side
1 2 3 next »