CCM 6.1 cluster (publisher and suscriber)

Unanswered Question
Dec 9th, 2008
User Badges:

hii, i have CCM v6.1 pubisher and another suscriber, i want to put the suscriber in a second site via a WAN connection, i wanna know if i have to verify some WAN consideration like 80ms of RTT and bandwidth, i read this in from this link:


http://www.cisco.com/en/US/docs/voice_ip_comm/cucm/srnd/6x/models_ps556_TSD_Products_Implementation_Design_Guide_Chapter.html#wp1043996

in the "Remote Failover Deployment Model"


the problem is that in this case the suscriber is an active ccm, than we must have wan considerations like bandwidth, because the publisher and the suscriber are sharing the applications.than there is a lot of communications between them.


but il my case, i think that the publisher is active ans the suscriber is working like backup(not active actualy).. please would i have the wan consideration in my case or not ? in the site they don't speak about a case where th suscriber is not active. it will be active only if the publisher connection lost. Thanks...


  • 1
  • 2
  • 3
  • 4
  • 5
Overall Rating: 5 (2 ratings)
Loading.
Jaime Valencia Tue, 12/09/2008 - 12:27
User Badges:
  • Cisco Employee,
  • Hall of Fame,

    2011

you need to meet the BW considerations (BW = 2xT1s) and RTT <= 80 ms as long as this is clustering over the WAN.


HTH


java


if this helps, please rate

a.rayen83 Tue, 12/09/2008 - 12:30
User Badges:

thank u, but in all the cases the suscriber is always active? because in my case i think that the susciber is only used for a backup, than why we need al this bandwidth and RTT?

Jaime Valencia Tue, 12/09/2008 - 12:32
User Badges:
  • Cisco Employee,
  • Hall of Fame,

    2011

for messages being exchanges (ICCS) between servers and DB replication. if you do not meet requirements we do not support the deployment


as long as it fails under clustering over the WAN you need to meet requirements.


HTH


java


if this helps, please rate


a.rayen83 Tue, 12/09/2008 - 12:37
User Badges:

okk, what do u man by 2xT1s? it's 2x1,54Mbs??? and this bandwidth is standrad for all the cases or we have a some procedures to calculate it?

Jaime Valencia Tue, 12/09/2008 - 12:40
User Badges:
  • Cisco Employee,
  • Hall of Fame,

    2011

it's in the doc you posted


For Unified CM 6.1 and later releases, a minimum of 1.544 Mbps (T1) bandwidth is required for Intra-Cluster Communication Signaling (ICCS) for 10,000 busy hour call attempts (BHCA) between sites that are clustered over the WAN. This is a minimum bandwidth requirement for call control traffic, and it applies to deployments where directory numbers are not shared between sites that are clustered over the WAN. The following equation may be used as a guideline to calculate the bandwidth for more than 10,000 BHCA between non-shared directory numbers at a specific delay:


Total Bandwidth (Mbps) = (Total BHCA/10,000) * (1 + 0.006 * Delay), where Delay = RTT delay in msec


This call control traffic is classified as priority traffic. Priority ICCS traffic is marked with IP Precedence 3 (DSCP 24 or PHB CS3).


•In addition to the bandwidth required for Intra-Cluster Communication Signaling (ICCS) traffic, a minimum of 1.544 Mbps (T1) bandwidth is required for database and other inter-server traffic for every subscriber server remote to the publisher in Unified CM 6.1 and later releases.


HTH


java


if this helps, please rate

allan.thomas Tue, 12/09/2008 - 12:40
User Badges:
  • Blue, 1500 points or more

Although the Subscriber as you mentioned will not be strictly active in the sense that IP phones are not registered to it, it will still be an active node in the cluster.


For example, regardless of whether there is no Intra-Cluster Communication Signalling (ICCS) between nodes across the WAN, you still need to provision bandwidth for database and other inter-server traffic for every subscriber.


This is stipulated with the SRND you have reference above, and the recommendation is for at least 1.5Mbps for this traffic.


Therefore although the subscriber is not active, it will still be replicating and communicating with the Publisher node. If this communications fails, then you may find all sorts of issues when IP Phones do finally register to this node.


Hope this helps.

Allan.

aliisa1976 Wed, 12/10/2008 - 12:14
User Badges:

Hi All,

I am doing the same implementation and I have few enquiries:

1) I have installed CCM 6.1(2) as publisher using MCS7845 hardware (IBM), I have one MCS7835 hardware (HP) and I want to use it as subscriber, during loading CCM application in the MCS7835 hardware I have got a message saying that Unity connection and CCM business edition is not supported by the hardware (MCS7835 from HP), could you please advise this setup is correct or not? can I use along with publisher business edition?


2)In the publisher, when I checked the license I saw only one node, could you please advise which license I have to order to have a total of 4 subscribers along with one publisher. I have two PAKs with 5,000 seats for MCS7845 hardware and I have also two PAKs for 2,500 seats for MCS7835 hardware, I want to know which one to use and which MAC address I have to use when I want to generate a license. I have 3 MCS7845 (IBM) and 2 MCS7835 (HP), I want to use one of the MCS7845 as publisher.


3) The publisher has two MAC addresses (2 NICs), please advise which MAC I have to use to generate the licenses.


4) What is the difference between the normal CCM edition and the business edition release 6.1(2)?


Thanks,

Ali Isa

Jaime Valencia Wed, 12/10/2008 - 12:59
User Badges:
  • Cisco Employee,
  • Hall of Fame,

    2011

CUCMBE is STANDALONE, NO SUBSCRIBERS.

it's all in one box solution.


provide primary MAC, ETH0.


BE is CUCM and CUC coresident, only 500 users.

read the CUCMBE section from SRND for full details on restrictions on CUCMBE deployments


HTH


java


if this helps, please rate



aliisa1976 Wed, 12/10/2008 - 20:15
User Badges:

Thanks Java for your reply, could you please answer me on the other enquiries that I have mentioned in my prevous message?


Thanks and regards,

Ali Isa


Jaime Valencia Thu, 12/11/2008 - 08:18
User Badges:
  • Cisco Employee,
  • Hall of Fame,

    2011

guess you didn't get the whole idea. if you can have only 1 server on CUCMBE you cannot order any more licenses. all other stuff has been answered


HTH


java


if this helps, please rate

aliisa1976 Thu, 12/11/2008 - 12:11
User Badges:

thanks Java,

I have call manager 6.1(2) NOT business edition. We have three MCS7845 (IBM) and two MCS7835 (HP) ordered separatly. We have received two 5,000 seats PAKs along with MCS 7845 and two 2,500 seats along with MCS7835 (HP). Could you please advise if I need one more 5,000 seats PAK for the MCS7845 hardware or not?


Do I need to apply all these licenses in the publisher or I have to apply each one in each hardware.


thanks,

Ali Isa


a.rayen83 Thu, 12/11/2008 - 08:02
User Badges:

hi Thomas, now in understand the notions of clustering, but i have a question, read this please, it is in the SRND callmanager 6.1

"The network should be engineered to provide sufficient prioritized bandwidth for all ICCS traffic, especially the priority ICCS traffic. Standard QoS mechanisms must be implemented to avoid congestion and packet loss. If packets are lost due to line errors or other "real world" conditions, the ICCS packet will be retransmitted because it uses the TCP protocol for reliable transmission. The retransmission might result in a call being delayed during setup, disconnect (teardown), or other supplementary services during the call. Some packet loss conditions could result in a lost call, but this scenario should be no more likely than errors occurring on a T1 or E1, which affect calls via a trunk to the PSTN/ISDN"

if i understand, when we have a call, the 2 callmanager are active?? because the signalisation is send frome the 1st CCM to the second!! so the 2 CCM are active? thank u

allan.thomas Thu, 12/11/2008 - 08:31
User Badges:
  • Blue, 1500 points or more

Ordinarily when you have a distributed cluster scenario, for example two CallManagers in different locations then both will be providing local call processing for that respective location. You would not necessarily register all your phone across the WAN to a single callmanager and then have them fail-over when their primary is down.


When a call attempt is made to an IP Phone from one location to the other then the local CCM to which the IP Phone is registered is responsible for the ICCS traffic across the WAN to the remote CallManager.


Thus you need to ensure that ICCS traffic receive the appropriate QoS bandwidth guarantees including database replication.


Regards

Allan.

Actions

This Discussion