Cisco Unified Communication Manager (Unified CM) co-loaded with Unified CCX on the same virtual machine (VM) or bare metal server is not supported.
Cisco Unified Intelligence Center is deployed co-loaded on the same virtual machine (VM) or bare metal server with Unified CCX and supports all the Unified CCX deployment models.
The following table depicts the deployment models that are supported in Unified CCX. These models have no bearing on which specific server model is used. The minimum server model required is identified by the Cisco Unified Communications Sizing Tool. This chapter provides general rules for design and considerations and limitations for each of these deployment models. This information allows an Unified CCX system planner or designer to understand what other similar deployment models are supported and to understand how to determine the best solution for a given set of requirements.
Table 1 Unified CCX Deployment Models
Unified CCX Deployment Model
Unified CCX Components on Server 1
Unified CCX Components on Server 2
Single-Server Non-High Availability Deployment Model—Unified CM Integration
Unified CCX deployment model integrated with Unifed CME is not supported in 9.0(1) and higher versions.
The following figure depicts the deployment when integrating Unified CCX with Unified CM. In this deployment, optional Unified CCX components shown with an asterisk (*) can be added. These components are:
Cisco Unified Work Force Management and Cisco Unified Quality Manager.
Cisco IM and Presence Server. For more details about deploying the Presence Server, refer the Cisco Unified Communications SRND, which is available at this URL: http://www.cisco.com/go/ucsrnd
Figure 1. Deployment Model of Unified CCX Integrated with Unified CM
ASR and TTS can be added in Unified CCX integrated with Unified CM. ASR and TTS software is not provided by Cisco. This software must be purchased from a vendor such a Nuance, Scansoft, or IBM. These vendors can provide design and server sizing requirements for their software.
When designing a Unified CCX deployment, the following rules apply:
When deploying for high availability (HA), the Unified CCX servers can be located in the same campus LAN to provide server redundancy. The Cisco Unified CCX servers can also be located in different sites separated by WAN to provide spatial redundancy.
For HA over LAN deployment, heartbeats are sent every half a second and failover occurs if 5 consecutive heartbeats are missed. For HA over WAN deployment, heartbeats are sent every second and failover occurs if missing 10 consecutive heartbeats. These values are not configurable.
You can locate the Unified Communications Manager servers that run CTI Managers with which Unified CCX communicates in the same campus LAN. In case of Unified CCX servers that are deployed over WAN, for better site redundancy, Cisco highly recommends that you deploy local Unified Communications Manager server at both sites.
If recording is going to be used for a high availability deployment, the Recording component must be redundant.
All agents for a Unified CCX deployment must be using phones that register to the same Unified CM cluster. Calls can be received from devices and callers on another Unified CM cluster (using inter-cluster trunks).
All Unified CCX deployments must be configured using the Cisco Unified Communications Sizing Tool to be supported. Only the MCS 7835 and 7845 models provide redundant power supplies, redundant fans, and redundant hot-swappable disk drives. Therefore, use the MCS 7835 or the MCS 7845 for higher resiliency.
High availability is supported with Cisco Unified IP-IVR but not when it is used as the queuing platform for Unified CCE.
Unified CCX software and database versions must be the same for both the master and standby nodes in a high availability deployment.
Different server models can be used in a high availability deployment with the following constraints.
The capacity of the subscriber HDD should be equal to or more than that of the publisher HDD.
In the case of different servers being used in a high availability deployment, the system capacity is determined by the smaller of the two servers.
Unified CCX solution works with a combination of software and hardware components, providing an open and flexible environment for customers to execute complex scripts, custom codes, documents, etc. Overloading any of the software/hardware components such as virtual memory, CPU, etc. could impact the solution performance. Cisco recommends reviewing and optimizing scripts, custom codes, documents etc., before they are loaded to the production setup; Cisco also recommends also constant monitoring of system component/hardware attributes like disk space and CPU utilization. For more details on optimization and best practices, please refer to "Cisco Unified Contact Center Express and IPIVR - Best Practices" which is available at: http://www.cisco.com/en/US/docs/voice_ip_comm/cust_contact/contact_center/crs/express_7_0/reference/guide/UCCX_Best_Practices.pdf
When deploying Quality Management and Workforce Management with Unified CCX, consider the following guidelines:
Quality Management and Workforce Management must be installed on separate servers from each other and Unified CCX. No form of co-residency is supported in this release with any other software, such as installing on Unified CCX or installing both Quality Management and Workforce Management on the same server.
WFO products do not provide redundancy. However, note that both Quality Management and Workforce Management do support redundant Unified CCX environment and are able to switch to a secondary or backup system upon failure of the primary Unified CCX system.
Unified CCX does not support the use of third party applications (for example, using TAPI) to control its devices.
The Single-Server Non-High Availability Deployment Model—Unified CM integration is for small deployments. This deployment model places a single instance of all four Unified CCX software components on the same server and uses Informix IDS as the database server.
This deployment model can support silent monitoring and recording for agents at any WAN-connected site by using desktop monitoring. (Refer the Cisco Unified CCX Software and Hardware Compatibility Guide for a list of phones that support desktop monitoring). It can also support SPAN port monitoring for agents on the VLAN segment local to Unified CCX server. This deployment model does not incorporate additional remote Monitoring components, so silent monitoring and recording is not possible for agents who are using the Cisco IP Phone Agent at remote sites. Similarly, silent monitoring and recording is not possible for agents at remote sites who are using phones that do not support desktop monitoring.
This deployment model allows the Unified CCX Engine to fail over to a backup CTI Manager if the primary CTI Manager fails. CTI ports and CTI route points should be grouped into device pools that have the same primary and secondary server list as those used for JTAPI communications with the CTI Managers.
Two-Server High Availability Deployment Model — Cisco Unified Communications Manager Integration
The Two-Server High Availability Deployment Model—Unified CM integration is for small to medium-sized contact centers requiring high availability. This deployment model incorporates redundant Unified CCX Engine, Database, Recording, and Monitoring components.
Informix IDS replication is used to keep the databases synchronized.
This deployment model can support silent monitoring and recording for agents at any WAN-connected site by using desktop monitoring. (Refer the Cisco Unified CCX Software and Hardware Compatibility Guide for a list of phones that support desktop monitoring). It can also support SPAN port monitoring for agents on the VLAN segment local to Unified CCX server. This deployment model provides redundancy for both recording and silent monitoring for all agents using desktop monitoring (regardless of location) or agents on the local VLAN using SPAN port monitoring. This deployment model does not incorporate additional remote Monitoring components, so silent monitoring and recording is not possible for agents who are using the Cisco IP Phone Agent at remote sites. Similarly, silent monitoring and recording is not possible for agents at remote sites who are using phones that do not support desktop monitoring.
This deployment model allows either Unified CCX Engine component to fail over to a backup CTI Manager if the primary server fails. CTI Ports and CTI Route Points should be grouped into device pools that have the same primary and secondary server list as that used for JTAPI communications to the CTI Managers.
In HA deployments, historical data comes from the database located in the standby engine node. On newer platforms, a higher number of historical reporting sessions during operating hours is supported for HA deployments. Refer the Cisco Unified Communications Sizing Tool for the limits.
Unified CCX supports high availability over WAN to provide site redundancy. In this deployment, the Unified CCX servers are located in two different sites across the WAN. Cisco recommends each site should have at least one Unified CM server that is running CTI Manager with which Unified CCX communicates. Agents and supervisors can be located in one of the sites where the Unified CCX server resides or in any other remote sites. The following figure depicts the deployment for Unified CCX high availability over WAN.
Figure 2. Unified CCX High Availability over WAN Deployment
The maximum allowed round-trip time (RTT) between Unified CCX servers is 80 ms.
Do not use the ping utility on the Unified CCX server to verify RTT as it will not provide an accurate result. The ping is sent as a best-effort tagged packet and is not transported using the same QoS-enabled path as the WAN traffic. Therefore, Cisco recommends that you verify the delay by using the closest network device to the Unified CCX servers, ideally the access switch to which the server is attached. Cisco IOS provides an extended ping capable of setting the Layer 3 type of service (ToS) bits to make sure the ping packet is sent on the same QoS-enabled path that the WAN traffic will traverse. The time recorded by the extended ping is the round-trip time (RTT), or the time it takes to traverse the communications path and return. Refer to the IOS document available at http://www.cisco.com/en/US/tech/tk365/technologies_tech_note09186a0080093f22.shtml#extend_ping for more detail.
Sufficient bandwidth must be provisioned for Unified CCX cluster, Unified CM cluster, remote agent/supervisor desktops and other optional components in order to deploy HA over WAN successfully.
Account for the bandwidth required for the following components:
Unified CCX Cluster and Unified CM Cluster Unified CCX cluster consumes bandwidth between the Unified CCX servers in high availability. If the Unified CM running CTI Manager that Unified CCX communicates with is remote, there would be additional bandwidth utilized by Unified CCX. Unified CM could consume significantly higher bandwidth for Intra-Cluster Communication Signaling (ICCS) between sites when deploying with Unified CCX. This is due to the additional number of call redirects and CTI/JTAPI communications encompassed in the intra-cluster communications. Unified CCX can be deployed as ACD to route and queue contacts for available agent or as IP-IVR to perform self-service. The bandwidth requirements for Unified CCX and Unified CM clusters are different depending on the deployment type.
The following table shows the minimum bandwidth requirement for Unified CCX and Unified CM clusters when deploying HA over WAN.
Table 2 Unified CCX HA over WAN Bandwidth Requirement
1 This column shows the database bandwidth required for Unified CM clustering over WAN and could be subject to change. For the final authorized value, refer to
Cisco Unified Communications Solution Reference Network Design (SRND) available at:
http://www.cisco.com/go/ucsrnd2 BHCA (Busy Hour Call Attempt) is the number of calls entering the system in the busy hour for Unified CCX or IP-IVR.
For Unified CCX Cluster in the preceding table:
The traffic between Unified CCX servers includes database replication, heartbeat and other communication between the Unified CCX HA servers.
The traffic between Unified CCX server and remote Unified CM server running CTI Manager is the JTAPI call signaling.
For Unified CM Cluster in the preceding table:
Database column includes traffic for database and other inter-server traffic for every Cisco Unified CM subscriber server remote to the Unified CM publisher.
ICCS column shows all the ICCS traffic between CallManager/CallManager services and CallManager/CTI Manager services running in the Unified CM nodes across sites.
As an example, assume the Unified CCX HA over WAN deployment has two sites and is used as ACD. Site 1 has the Unified CCX, one Unified CM publisher and two Unified CM subscribers. Site 2 has the other Unified CCX and two Unified CM subscribers. Unified CCX in site 1 communicates with Unified CM subscriber in site 2 for JTAPI signaling. In the busy hour, there are 1500 calls coming into Unified CCX that get routed or queued for agents.
The maximum supported response time between the Unified CCX server and the Microsoft Exchange server is 80 ms.
For Unified CCX cluster, bandwidth required is: 1.2 Mbps + 800 kbps (0.8 Mbps) = 2 Mbps For Cisco Unified CM cluster, there are two Unified CM subscribers remote from the Unified CM publisher and the BHCA is 1500. Bandwidth required is: 1.544 Mbps × 2 + 70 kbps × 15 (1.05 Mbps) = 4.138 Mbps In total, 6.138 Mbps between sites is required for this deployment.
Agents and Supervisors In HA over WAN deployment, agents and supervisors could reside in either Unified CCX sites or they could be remote depending on the location of active Unified CCX server at the time of operation. Bandwidth should be provisioned for remote agents between sites using the maximum number of agents from the two sites. Estimate the required bandwidth using the Cisco Agent Desktop Bandwidth Calculator available at: http://www.cisco.com/en/US/products/sw/custcosw/ps427/prod_technical_reference_list.html
Customers might have the following optional components deployed across the WAN from Unified CCX or Unified IP IVR. Ensure to account for the additional bandwidth required in their HA over WAN deployment.
Wallboard Server: Figure out the amount of data that is pulled from Unified CCX database to the remote wallboard server.
Enterprise Database: Estimate the total amount of data that is retrieved through the database steps from the remote enterprise database.
SMTP Server: If the SMTP server is remote from the Unified IP IVR, find out the average size of each outgoing email and calculate the total.
Quality of Service
Quality of service (QoS) must be enabled and engineered correctly on the network to provide consistent and predictable end-to-end levels of service. Unified CCX software does not mark any network packet, thus ensure that you mark the traffic at the network edge routers.
The following table shows the recommendations on the QoS markings for Unified CCX HA over WAN deployment.
Table 3 QoS Rrecommendations for Unified CCX HA over WAN
3 The database traffic may be re-prioritized to a higher priority data service (for example, IP Precedence 2 [DSCP 18 or PHB AF21] if required by the particular business needs). An example of this is the usage of outbound dialer in Unified CCX, which relies on writing data to the Config Datastore.
Consider the following when deploying High Availability over WAN with Unified CCX:
Deploy ASR/TTS server locally in each Unified CCX site
Set up Unified CCX to use the local Unified CM servers for both primary and secondary in the following configurations. If this is not possible, at least the primary Unified CM server should be local.
AXL Service Provider
JTAPI Provider for Unified CM Telephony Subsystem
JTAPI Provider for Resource Manager/Contact Manager Subsystem
There will be significant delays in agent login during Unified CCX failover if AXL and JTAPI communications are made over the WAN, especially under load conditions.
Assign the two sets of CTI Port (one for the master and other for the standby engine) to different device pools, regions and locations, etc., in the CTI Port Group
Data in Agent Datastore, Historical Datastore and Repository Datastore of Informix IDS database start merging after the network partition is restored and this could potentially generate heavy data traffic over the WAN. Cisco recommends restoring the WAN link during after hours to minimize the performance impact.
Do not support VPN tunneling across the WAN.
Unified CCX Web Chat
As part of the Premium offering, Unified CCX agents can service customer chat requests using the Agent Web Chat Application from the Cisco Agent Desktop or through a standalone browser.
As stated in the overview section, this feature requires a Cisco SocialMiner deployment to accept and relay the contact requests from a customer website. One SocialMiner deployment can serve only one Unified CCX deployment (single node or High Availability deployment).
Deployment Scenario 1: Customer Website in Demilitarized Zone (DMZ)
Figure 3. Customer Website in DMZ
The Unified CCX is deployed inside the enterprise firewall and SocialMiner is deployed inside company premises in the DMZ zone along with the customer website. The DMZ zone is open to all HTTP/HTTPS traffic from the Internet. The Unified CCX is shielded from all outside traffic except the traffic coming from the DMZ zone. Even from the DMZ zone the Unified CCX only responds on HTTP/HTTPS and BOSH (7443/7071) ports.
Deployment Scenario 2: Customer Website in Public Cloud/Domain
Figure 4. Customer Website in Public Cloud/Domain
The Unified CCX is deployed inside the enterprise firewall and SocialMiner is deployed inside company premises in the DMZ. The customer website is on a public cloud or public domain and is open to the Internet. The DMZ zone is open to all HTTP/HTTPS traffic from the Internet. The Unified CCX is shielded from all outside traffic except the traffic coming from the DMZ zone. Even from the DMZ zone the Unified CCX only responds on HTTP/HTTPS and BOSH (7443/7071) ports.
One variation of the preceding scenario can be an addition of a proxy server that can intercept and relay all calls going to SocialMiner.
Unified CCX High Availability Deployment
Web Chat supports failover and failback as per the feature description table earlier. You should follow the same considerations and design strategy, as the one considered for the overall Unified CCX high availability deployment.
Other Design Considerations
Consider the following when designing your Unified CCX system:
High availability requires additional disk space, so historical call reporting capacity may be reduced. Historical call reporting capacity also depends upon BHCC, hours of operation per day, and days of operation per week.
G.711 call recording requires about 1MB per minute. G.729 call recording requires about 256KB per minute.
The following categories of data use hard disk space:
Linux Server OS, Unified CCX Software, and Informix Database Management Software
Unified CCX Logs
The Unified CCX Database (comprised of 4 data stores)
Recording Files. Systems planners and designers should attempt to estimate the impact of each in order to determine hard disk requirements. The Cisco Customer Response Solutions Installation Guide provides more information about disk size requirements for very large installations.
The Unified CM sizing tools assume devices are evenly distributed across all servers. CTI Route Points is configured as part of a device pool that homes primarily to the same Cisco Unified Communications Manager Server as the primary CTI Manager being used; it may be required to run the Cisco Unified Communications Manager sizing tool on a per location or per server basis.
The Unified CM QSIG path replacement feature is not supported for Unified CCX calls. For additional information about Unsupported Features in Unified CM, see the current release notes for Unified CCX.
Unified CM Forced Authorization Codes and Client Matter Codes should be turned off for all route patterns in the Unified CM cluster that are used by Unified CCX. Enabling these features for route patterns that aren't used by Unified CCX does not affect Unified CCX.
For a list of unsupported features in Unified CM with Unified CCX, refer to the current release notes for Unified CCX.
CAD supports only one localized language per Unified CCX cluster which is determined at installation. As a result, all Cisco Agent and Supervisor Desktop applications in that Unified CCX cluster must use the same language.
An agent can log in using Extension Mobility but the agent phone must be in the Unified CM cluster that is used by Unified CCX.
Sometimes new releases of Unified CM will not support Unified CCX immediately at Unified CM first customer ship (FCS) time. Some organizations may be early adopters of new Unified CM releases and may be slowed from migrating to new Unified CM releases and using new Unified CM features if Unified CCX is installed with that same Unified CM cluster. Therefore, in some scenarios, it may make sense to have a separate Unified CM cluster for Unified CCX.
Cisco Jabber runs in two modes: Deskphone Mode and Softphone Mode. Unified CCX only supports Cisco Jabber as an agent device in Softphone Mode.
Cisco Agent Desktop/Cisco Supervisor Desktop does not support video operation if you are using Cisco Jabber for Windows as agent phone.
The following features are not supported if you are using Cisco Jabber for Windows as agent phone:
Multiple Cisco Unified CCX Clusters integrated with a single Cisco Unified Communications Manager cluster
It is possible to integrate multiple Unified CCX clusters with a single Cisco Unified Communications Manager cluster.
There is no limit on the number of Unified CCX clusters supported with a single Unified CM cluster as long as the combined agent phones, CTI ports, CTI route points, etc., utilized by all Unified CCX clusters were used to size Unified CM.
To determine if you need more than one CTI Manager, refer to the Cisco Unified Communications Solution Reference Network Design (SRND), available at http://www.cisco.com/go/ucsrnd. If your deployment requires more than one CTI Manager, Cisco recommends that you load-balance Unified CCX and other CTI applications across the various CTI Managers in the cluster to provide maximum resilience, performance, and redundancy. For additional CTI Manager best practices, refer to the Cisco Unified Communications Solution Reference Network Design (SRND), available at http://www.cisco.com/go/ucsrnd.
If more than one primary subscriber is required to support your configuration, distribute all agents equally among the subscriber nodes. This configuration assumes that the busy-hour call attempts (BHCA) is uniform across all agents.
Each Unified CCX cluster is standalone and independent from other Unified CCX clusters. There is no communication or synchronization between the Unified CCX clusters. Agents should operate in can only one Unified CCX cluster. Unified CM Telephony Triggers (CTI Route Points) and CTI ports should be different across Unified CCX clusters.
In the list of Resources in Unified CCX Administration, each Unified CCX cluster displays all the agents in the Cisco Unified Communications Manager cluster, even though the agents can operate and log in to one Unified CCX cluster. This situation requires that the Unified CCX Administrator be aware of which resources are associated with each cluster. The Unified CCX Administrator can mitigate this situation by having a unique naming convention for resources associated with a particular Unified CCX cluster.
This deployment is not intended to provide Unified CCX redundancy across different Unified CCX clusters. If a Unified CCX cluster fails, the agents that operate in this cluster cannot operate in other Unified CCX clusters. If another Unified CCX cluster is configured to accept the calls that were originally sent to the Unified CCX cluster that failed, there will be no report integration between the Unified CCX clusters.
This deployment does not change the characteristics and design recommendations of each individual Unified CCX cluster. For example, within a Unified CCX cluster, high availability is still supported.
Virtualized Deployment on Cisco Unified Computing System (UCS)
Unified CCX can be deployed as an application on a virtual machine on the VMWare platform running on the Cisco Unified Computing System (UCS) hardware. This allows Cisco Unified Contact Center Express to be part of a virtualized deployment on the same hardware along with other Cisco Unified Communications applications with resulting savings due to server consolidation and reduced total cost of ownership (TCO).
Unified CCX is installable on a virtual machine using a standard Open Virtualization Format Template (called OVA template). Using a single OVA templates we can install 100, 300, and 400 agents profile. Each OVA template is associated with a certain number of CPU and specific main memory and hard disk capacity, and supports a predefined set of scalability parameters (including the number of agents). Please refer to the Unified CCX Data Sheet for information on scalability parameters supported by each OVA template.
For guidelines on deployment of Unified Communications (UC) applications (including Unified CCX) on UCS servers, refer to the "Deploying Unified Communications on Virtualized Servers" section in the Cisco Unified Communications System SRND available at the following location:
For more details on the hardware requirements, VMWare requirements, supported VMW are features, virtual machine sizing and best practices for this deployment, please refer to the following link: http://www.cisco.com/go/uc-virtualized
Cisco HCS Deployment Models
Unified Contact Center Express supports both on-premises and hosted deployment models in Cisco HCS to provide contact center capabilities to customers. Contact your account team to obtain a copy of the Cisco HCS Design and Implementation Guide to configure and use the Unified Contact Center Express in Cisco HCS.
The following limitations apply for HCS deployment of Unified CCX.
Table 7 Deployment Limitations for HCS
Does not support Work Flow Management (WFM), Quality Management (QM/AQM), and Compliance Recording (CR) applications.
The maximum WAN round trip delay that Unified CCX supports remains the same when deployed in HCS. The WAN between Unified CM and Unified CCX, or the HCS customer premises to data center over VPN, should meet these requirements.