Cisco IP Contact Center Enterprise Edition Releases 5.0 and 6.0 Solution Reference Network Design (SRND)
Sizing IPCC Components and Servers
Downloads: This chapterpdf (PDF - 296.0KB) The complete bookPDF (PDF - 3.26MB) | Feedback

Sizing IPCC Components and Servers

Table Of Contents

Sizing IPCC Components and Servers

Sizing Considerations for IPCC

Core IPCC Components

Minimum Hardware Configurations for IPCC Core Components

Additional Sizing Factors

Peripheral Gateway and Server Options

CTI OS

Cisco Agent Desktop Component Sizing

Cisco Agent Desktop Base Services

Cisco Agent Desktop VoIP Monitor Service

Cisco Agent Desktop Recording and Playback Service

Summary


Sizing IPCC Components and Servers


Proper sizing of your Cisco IP Contact Center (IPCC) Enterprise solution is important for optimum system performance and scalability. Sizing considerations include the number of agents the solution can support, the maximum busy hour call attempts (BHCA), and other variables that affect the number, type, and configuration of servers required to support the deployment. Regardless of the deployment model chosen, IPCC Enterprise is based on a highly distributed architecture, and questions about capacity, performance, and scalability apply to each element within the solution as well as to the overall solution.

This chapter presents best design practices focusing on scalability and capacity for IPCC Enterprise deployments. The design considerations, best practices, and capacities presented in this chapter are derived primarily from testing and, in other cases, extrapolated test data. This information is intended to enable you to size and provision IPCC solutions appropriately.

Sizing Considerations for IPCC

This section discusses the following IPCC sizing considerations:

Core IPCC Components

Minimum Hardware Configurations for IPCC Core Components

Additional Sizing Factors

Core IPCC Components

When sizing IPCC deployments, IP Telephony components are a critical factor in capacity planning. Good design, including multiple Cisco CallManagers and clusters, must be utilized to support significant call loads. For additional information on Cisco CallManager capacity and sizing of IP Telephony components, refer to Sizing Cisco CallManager Servers For IPCC, and to the Cisco IP Telephony Solution Reference Network Design (SRND) guide, available at

http://www.cisco.com/go/srnd

Additionally, because of varying agent and skill group capacities, proper sizing of the CTI OS and Cisco Agent Desktop servers should be considered together with the IP Telephony components.

Finally, the remaining ICM components, while able to scale extremely well, are affected by specific configuration element sizing variables that also have an impact on the system resources.

These factors, discussed in this section, must be considered and included in the planning of any deployment.

The information presented in Figure 5-1, Figure 5-2, and Table 5-1 does not apply equally to all implementations of IPCC. The data is based on testing in particular scenarios, and it serves only as a guide, along with the sizing variables information in this chapter. As always, you should be conservative when sizing and should plan for growth.


Note Sizing considerations are based upon capacity and scalability test data. Major ICM software processes were run on individual servers to measure their specific CPU and memory usage and other internal system resources. Reasonable extrapolations were used to derive capacities for co-resident software processes and multiple CPU servers. This information is meant as a guide for determining when ICM software processes can be co-resident within a single server and when certain processes need their own dedicated server. Table 5-1 assumes that the deployment scenario includes two fully redundant servers that are deployed as a duplexed pair. While a non-redundant deployment might theoretically deliver higher capacity, no independent testing has been done to validate this theory. Therefore, you can and should refer to Table 5-1 for sizing information about simplexed as well as duplexed deployments.



Note The Cisco IP Contact Center solution does not provide a quad-processor Media Convergence Server (MCS) at this time. If extra performance is required beyond the limits described in the table below, it might be possible to use an off-the-shelf quad-processor server in lieu of the MCS 7845. For server specifications, refer to the Cisco Intelligent Contact Management Software Bill of Materials (BOM) documentation available at http://www.cisco.com/univercd/cc/td/doc/product/icm/ccbubom/index.htm.


The following notes apply to all figures and tables in this chapter:

The number of agents indicates the number of logged-in agents.

Server types:

APG = Agent Peripheral Gateway

CAD = Cisco Agent Desktop

HDS = Historical Data Server

PRG = Progger

RGR = Rogger

Figure 5-1 Minimum Servers Required for Typical IPCC Deployment with CTI Desktop

The following notes apply to Figure 5-1:

Sizing is based upon the Cisco MCS 7845 (3.0 GHz or higher) and 5 skill groups per agent.

Voice Response Unit (VRU) and Cisco CallManager components are not shown.

For more than 2,000 agents, refer to Table 5-1.

The Agent Peripheral Gateway (APG) consists of a Generic PG (Cisco CallManager PIM and VRU PIM), CTI Server, and CTI OS.

For more information about APG deployment and configuration options, see Figure 5-3 and Figure 5-4.

Figure 5-2 Minimum Servers Required for Typical IPCC Deployment with Cisco Agent Desktop

The following notes apply to Figure 5-2:

Sizing is based upon the Cisco MCS 7845 (3.0 GHz or higher) and 5 skill groups per agent.

Voice Response Unit (VRU) and Cisco CallManager components are not shown.

For more than 2,400 agents, refer to Table 5-1.

The Agent Peripheral Gateway (APG) consists of a Generic PG (Cisco CallManager PIM and VRU PIM), CTI Server, and CTI OS.

Table 5-1 Sizing Information for IPCC Components and Servers 

Component
Server Model 1
Number of CPUs
Maximum Agents
Notes

Progger:

Peripheral Gateway, Router, and Logger

MCS-7835I-3.0-CC1

MCS-7835H-3.0-CC1

 

MCS-7845H-3.0-CC1

 

1

 

 

2

 

100

 

250

Cannot be co-resident with Administrative Workstation (AW) or Historical Data Server (HDS).

Maximum of 50 simultaneous queued calls.

Maximum of 125 simultaneous queued calls.

Logger database is limited to 14 days.

Maximum of 100 agents on MCS-7845 if using a co-resident Cisco Agent Desktop server.

Maximum of 50 agents on MCS-7845 if using a co-resident Dialer.

Maximum of 25 agents on MCS-7845 if using a co-resident Cisco Agent Desktop server and Dialer.

Outbound Dialer is not supported on the MCS-7835 in the Progger configuration.

Rogger:

Router and Logger

MCS-7835I-3.0-CC1

MCS-7835H-3.0-CC1

MCS-7845H-3.0-CC1

 

1

2

 

500

1,500

 

Router

MCS-7845H-3.0-CC1

Third-party quad server

2

4

5,000

6000

MCS-7835 not supported

Logger

MCS-7845H-3.0-CC1

Third-party quad server

2

4

5,000

6,000

MCS-7835 not supported

Administrative Workstation (AW) and Historical Data Server (HDS)

MCS-7835I-3.0-CC1

MCS-7835H-3.0-CC1

MCS-7845H-3.0-CC1

Third-party quad server

 

1

2

4

 

500

5,000

6,000

AW/HDS cannot be co-resident with a Progger, Rogger, Router, Logger, or PG.

Maximum of 2 AW/HDS supported with a single Logger; maximum of 4 with duplexed Loggers.

WebView server can be co-resident with HDS for up to 50 simultaneous users.

WebView Reporting Server

MCS-7835I-3.0-CC1

MCS-7835H-3.0-CC1

MCS-7845H-3.0-CC1

 

1

2

50 simultaneous WebView clients

A total of 4 WebView servers may be deployed to reach 200 simultaneous WebView clients.

Difference between MCS-7845 and MCS-7835 is the number of agents supported by AW/HDS.

Agent PG

(Inbound only)

MCS-7835I-3.0-CC1

MCS-7835H-3.0-CC1

 

MCS-7845H-3.0-CC1

 

1

 

2

 

250

 

500

Refer to Figure 5-3 and Figure 5-4 for more details about Agent PG configuration options.

Up to 150 Cisco Agent Desktop agents are supported on an MCS-7835 server Agent PG.

Up to 300 Cisco Agent Desktop agents are supported on a MCS-7845 server Agent PG.

VRU ports should not exceed half of the maximum supported agents listed in the Maximum Agents column. Additional VRU PGs can be deployed to accommodate a greater number of VRU ports.

For more information on the various Agent PG deployment options, see Peripheral Gateway and Server Options.

Voice Response Unit (VRU) PG

MCS-7835I-3.0-CC1

MCS-7835H-3.0-CC1

MCS-7845H-3.0-CC1

 

1

2

 

600 ports

1200 ports

Use the number of ports instead of agent count.

Average of 5 Run VRU Script Nodes per call.

Maximum of 8 PIMs per MCS-7845 and 4 PIMs per MCS-7835. Not to exceed 2 PIMs per 300 ports on a Generic PG.

Agent PG with Outbound Voice (includes Dialer and Media Routing PG)

MCS-7835H-3.0-CC1

MCS-7845H-3.0-CC1

1

2

(Inbound agents) + (2.5 * Outbound agents) < 250

(Inbound agents) + (2.5 * Outbound agents) < 500

Moving the dialer off of the Agent PG has no effect on the total number of Outbound agents supported.

Each transfer to a VRU port is equivalent to an agent.

Dialer only

MCS-7835H-3.0-CC1

MCS-7845H-3.0-CC1

1

2

100

200

Each transfer to a VRU port is equivalent to an agent.

Agent PG with Media Blender (Collaboration includes Media Routing PG)

MCS-7835H-3.0-CC1

MCS-7845H-3.0-CC1

1

2

Up to 500 total sessions

Media Routing (MR) PG co-residency requires the MCS-7845. See subsequent rows of this table for capacity numbers.

Media Blender

MCS-7845H-3.0-CC1

2

Up to 500 total sessions

MCS-7835 is not supported.

With MCS-7845:

Single-session chat: 250 agents and 250 callers

Blended collaboration: 250 agents and 250 callers

Multi-session chat: 125 agents and 375 callers

Web Collaboration Server

MCS-7845H-3.0-CC1

2

500 total sessions or 250 one-to-one

MCS-7835 is not supported.

With MCS-7845:

Single-session chat: 250 agents and 250 callers

Blended collaboration: 250 agents and 250 callers

Multi-session chat: 125 agents and 375 callers

Dynamic Content Adapter (DCA) for Web Option

MCS-7845H-3.0-CC1

2

100

MCS-7835 is not supported.

DCA co-residency is not supported.

Overall limitation (MCS-7845): 100 concurrent DCA sessions.

Email Manager Server

MCS-7835H-3.0-CC1

MCS-7845H-3.0-CC1

1

2

See note.

1000 (max)

MCS-7835 is not supported.

Less than 10 agents: all Cisco Email Manager components and databases co-exist on single server (MCS-7845).

Up to 250 agents: 2 servers - Cisco Email Manager AppServer, UI Server, and WebView on first; database server (Primary, LAMBDA, and CIR databases) on second.

Up to 500 agents: 4 servers - Cisco Email Manager AppServer on first; Cisco Email Manager UI Server (first) and WebView server on second; Cisco Email Manager UI Server (second) on third; database server on fourth. (In this scenario, an MCS-7835 may be used for the second UI Server box.)

Up to 1000 agents: 7 servers - Cisco Email Manager AppServer on first (quad processor recommended); Cisco Email Manager UI Server (first) and WebView server on second; Cisco Email Manager UI Server (second) on third; Cisco Email Manager UI Server (third) on fourth; Cisco Email Manager UI Server (forth, required if more than 750 agents) on fifth, database server (Primary and LAMDA) on sixth; database server (CIR) on seventh. (In this scenario, an MCS-7835 may be used for the n+1 UI Server boxes.)

For sizing information, refer to the Cisco Intelligent Contact Management Software Bill of Materials (BOM) documentation available at http://www.cisco.com/en/US/partner/products/sw/custcosw/ps1001/products_usage_guidelines_list.html.

Internet Service Node (ISN) Application Server and Voice Browser

 

 

 

For the server specifications for the Internet Service Node (ISN), refer to the Cisco Internet Service Node (ISN) Software Bill of Materials available at http://www.cisco.com/univercd/cc/td/doc/product/icm/ccbubom/index.htm.

IP IVR Server

 

 

 

For the IP IVR server specifications, refer to the Cisco IPCC Express and IP IVR Configuration and Ordering Tool, available at

http://www.cisco.com/en/US/partner/products/sw/custcosw/ps1846/prod_how_to_order.html

1 The MCS-7835I-3.0-CC1 and MCS-7835H-3.0-CC1 servers are no longer available from Cisco, but they can still be used in an IPCC configuration. Cisco currently sells the MCS-7835I1-3.0-CC1 server as a replacement for the MCS-7835I-3.0-CC1 and the MCS-7835H1-3.0-CC1 as a replacement for the MCS-7835H-3.0-CC1.


Minimum Hardware Configurations for IPCC Core Components

The sizing information in Table 5-1 must be applied to each deployment to gauge server sizing requirements accurately. The following considerations and guidelines apply to the information in Table 5-1:

Formal and critical call center deployments are encouraged to use dual CPU server configurations, especially for the Progger.

You must adhere to the capacity limits for the number of agents on each system component. Agent capacities are based upon a maximum of 30 BHCA per agent and 90 calls per IVR port. Adjustments either way are not supported: fewer agents does not necessarily equate to higher BHCA, nor does lower BHCA equate to a greater number of agents supported.

It is possible to use different server model configurations for the Rogger and the PGs, as long as you observe the BHCA and agent maximums for each separate component and adhere to the recommendations in the Cisco Intelligent Contact Management Software Bill of Materials (BOM) documentation.

The capacity maximums assume a normal amount of CTI traffic for each given configuration. Extraordinary CTI traffic (from very large IVRs, for example) will decrease the BHCA and agent maximums.

The capacity numbers are based on an average of 5 Run Voice Response Unit (VRU) scripts, running consecutively in the ICM script, per IVR call. If a deployment has a more complex ICM/IVR script than this, it will also decrease the maximum BHCA and agent capacity.

The capacity numbers are also based on 5 skill group per agent. If a deployment has more than five groups per agent, it will also decrease the maximum BHCA and agent capacity and should be handled on a case-by-case basis.

Additional Sizing Factors

Many variables in the IPCC configuration and deployment options can affect the hardware requirements and capacities. This section describes the major sizing variables and how they affect the capacity of the various IPCC components. In addition, Table 5-2 summarizes the sizing variables and their effects.

Busy Hour Call Attempts (BHCA)

The number of calls attempted during a busy hour is an important metric. As BHCA increases, there is an increase in the load on all IPCC components, most notably on Cisco CallManager, IP IVR, and the Cisco CallManager PG. The capacity numbers for agents assume up to 30 calls per hour per agent.

Agents

The number of agents is another important metric that will impact performance of most IPCC server components including Cisco CallManager clusters. For impact of agents on the performance of Cisco CallManager components, see Sizing Cisco CallManager Servers For IPCC

Skill Groups

The number of skill groups per agent has significant effects on the CTI OS Server, the Cisco CallManager PG, and the ICM Router and Logger. Cisco recommends that you limit the number of skill groups per agent to 5 or fewer, when possible, and that you periodically remove unused skill groups so that they do not affect system performance. You can also manage the effects on the CTI OS server by increasing the value for the frequency of statistical updates.

Queuing

The IP IVR places calls in a queue and plays announcements until an agent answers the call. For sizing purposes, it is important to know whether the IVR will handle all calls initially (call treatment) and direct the callers to agents after a short queuing period, or whether the agents will handle calls immediately and the IVR will queue only unanswered calls when all agents are busy. The answer to this question determines very different IVR sizing requirements and affects the performance of the ICM Router/Logger and Voice Response Unit (VRU) PG. Required VRU ports can be determined using the Cisco IPC Resource Calculator. (See Cisco IPC Resource Calculator, for more information.)

ICM Script Complexity

As the complexity and/or number of ICM scripts increase, the processor and memory overhead on the ICM Router and VRU PG will increase significantly. The delay time between replaying Run VRU scripts also has an impact.

Reporting

Real-time reporting can have a significant effect on Logger, Progger, and Rogger processing due to database access. A separate server is required for an Administrative Workstation (AW) and/or Historical Data Server (HDS) to off-load reporting overhead from the Logger, Progger, and Rogger.

IVR Script Complexity

As IVR script complexity increases with features such as database queries, the load placed upon the IVR server and the Router also increases. There is no good rule of thumb or benchmark to characterize the IP IVR performance when used for complex scripting, complex database queries, or transaction-based usage. Cisco recommends that you test complex IVR configurations in a lab or pilot deployment to determine the response time of database queries under various BHCA and how they affect the processor and memory for the IVR server, PG, and Router.

IP IVR Self-Service Applications

In deployments where the IP IVR is also used for self-service applications, the self-service applications are in addition to the IPCC load and must be factored into the sizing requirements as stated in Table 5-1.

Third-Party Database and Cisco Resource Manager Connectivity

Carefully examine connectivity of any IPCC solution component to an external device and/or software to determine the overall effect on the solution. Cisco IPCC solutions are very flexible and customizable, but they can also be complex. Contact centers are often mission-critical, revenue-generating, and customer-facing operations. Therefore, Cisco recommends that you engage a Cisco partner (or Cisco Advanced Services) with the appropriate experience and certifications to help you design your IPCC solution.

Extended Call Context (ECC)

The ECC usage impacts PG, Router, Logger, and network bandwidth. There are many ways that ECC can be configured and used. The capacity impact will vary based on ECC configuration and should be handled on a case-by-case basis.

Peripheral Gateway and Server Options

An ICM Peripheral Gateway (PG) translates messages coming from the Cisco CallManager servers, the IP IVR, or other third-party automatic call distributors (ACDs) or voice response units (VRUs) into common internally formatted messages that are then sent to and understood by the ICM. In the reverse, it also translates ICM messages so that they can be sent to and understood by the peripheral devices.

The Peripheral Interface Manager (PIM) is the software process that runs on the PG and performs the message translation and control. Every peripheral device that is part of the IPCC solution must be connected to a PG and PIM.

Figure 5-3 and Figure 5-4 illustrate various configuration options for the Agent PG with CTI OS and Cisco Agent Desktop.

Table 5-2 lists PG and PIM sizing recommendations

Figure 5-3 Agent PG Configuration Options with CTI OS

Figure 5-4 Agent PG Configuration Options with Cisco Agent Desktop

Table 5-2 PG and PIM Sizing Recommendations 

Sizing Variable
Recommendation, Based on ICM Software Release 5.0

Maximum number of PGs per ICM

80

Maximum PG types per server platform

Up to 2 PG types are permitted per server, provided that any given server is limited to the maximum agent and VRU port limitations outlined in Table 5-1.

Can PGs be remote from ICM?

Yes

Can PGs be remote from Cisco CallManager or IP IVR?

No

PIM types

Cisco CallManager, IVR, Media Routing (MR), and ACD

Maximum number of PIMs per PG

Actual number of IVR PIMs is determined by the size of the IPCC deployment (agents, IVR ports, BHCA, and so forth). Under most circumstances, 5 PIMs per PG (Agent PG) and 8 PIMs per PG (standalone PG) is a reasonable limit.

Maximum number of PIM types per PG (CTI server may be added)

2 + CTI Server

Maximum number of IVRs controlled by one Cisco CallManager

Refer to the Cisco IP Telephony Solution Reference Network Design (SRND), available at http://www.cisco.com/go/srnd.

Maximum number of CTI servers per PG

1

Can PG be co-resident with Cisco CallManager on Media Convergence Server (MCS)?

No


CTI OS

The CTI OS is most commonly configured as a co-resident component on the Agent PG (see Figure 5-3 and Figure 5-4), supporting up to 500 agents.

Table 5-3 lists additional sizing factors for CTI OS.

Table 5-3 CTI OS Sizing Factors 

Sizing Factor
MSC-7845
MSC-7835
Comments

Maximum number of agents plus supervisors

500

250

 

Maximum number of supervisors

50

25

10% of supported agents1

Maximum number of teams

50

25

10% of supported agents1

Maximum BHCA per agent

30

30

Supervisors do not receive new calls.

Maximum number of agents plus supervisors per team

100

50

20% of supported agents1

Maximum number of supervisors per team

10

5

2% of supported agents1

Maximum number of agents per supervisor

100

50

20% of supported agents1

Skill groups per agent

5

5

 

Extended Call Context (ECC)

None

None

 

1 These percentages apply to the CTI OS Server agent capacity and not to the entire Contact Center capacity.


The numbers in Table 5-3 are based on the following assumptions:

Hyper-threaded is enabled for the MCS server.

The traffic profile is:

85% of calls answered by agents

10% of calls transferred

5% of calls conferenced

For more than 500 CTI agents, additional Agent PG instances are off-loaded (500 agents each) to additional servers. The CTI OS capacity decreases when CTI OS configuration values differ from those listed in Table 5-3. For example, the CTI OS capacity decreases as the number of skill groups per agent increases. In this case, platform resource usage increase significantly due to the increase in work load required to query and update agents and skill groups.

Cisco Agent Desktop Component Sizing

For details on the components and architecture of the Cisco Agent Desktop, see Agent Desktop and Supervisor Desktop.

Server capacities for the Cisco Agent Desktop CTI Option vary based on the total number of agents, whether or not Switched Port Analyzer (SPAN) monitoring and recording is used, and the number of simultaneous recordings.

This section presents sizing guidelines for the following installable Cisco Agent Desktop Server components:

Cisco Agent Desktop Base Services

Cisco Agent Desktop VoIP Monitor Service

Cisco Agent Desktop Recording and Playback Service

Cisco Agent Desktop Base Services

The Cisco Agent Desktop Base Services consist of a set of application servers that run as Microsoft Windows 2000 services. They include Chat Service, Directory Services, Enterprise Service, IP Phone Agent Service, LDAP Monitor Service, Licensing and Resource Manager Service, Recording and Statistics Service, and Sync Service. In addition, there are application servers that may be placed on the same or separate computers as the Base Servers. These additional applications include the VoIP Monitor Service and the Recording and Playback Service.

A set of Cisco Agent Desktop Base Services plus the additional application servers correspond to a logical call center (LCC). Table 5-4 lists the maximum number of agents that a single LCC can support for various sizes of enterprises.

Table 5-4 Maximum Number of Agents Supported by a Logical Call Center (LCC) 

Enterprise Size
Desktop Agents Only
IP Phone Agents Only
Mixed

Small

100

50

33 of each

Medium

300

150

100 of each

Large (multiple PG and CTI Servers)

500

250

333 of each

Maximum LCC capacity with multiple PG and CTI Servers

1000

500

333 of each


Cisco Agent Desktop VoIP Monitor Service

The VoIP Monitor Service enables the silent monitoring and recording features. For Desktop Monitoring, the VoIP Monitor Service has no impact on design guidance for Agent PG scalability. When using Switched Port Analyzer (SPAN) monitoring, the VoIP Monitor Service may be co-located on the Agent PG for up to 100 agents. When Remote Switched Port Analyzer (RSPAN) monitoring and recording are required for more than 400 agents, the VoIP Monitor Service must be deployed on a dedicated server (an MCS-7835 server or equivalent). Each dedicated VoIP Monitor Service can support up to 400 agents.

Cisco Agent Desktop Recording and Playback Service

The Recording and Playback Service stores the recorded conversations and makes them available to the Supervisor Log Viewer application.

A co-resident Recording and Playback Service can support up to 32 simultaneous recordings. A dedicated Recording and Playback Service (which is available in the Premium offering) can support up to 80 simultaneous recordings. The capacity of the Recording and Playback Service is not dependent on the codec that is used.

Table 5-5 summarizes the raw Recording and Playback Service capacity.

Table 5-5 Capacity of Recording and Playback Service 

Recording and Playback Service Type
Maximum Simultaneous Recordings

Co-resident

32

Dedicated

80


Summary

Proper sizing of IPCC components requires analysis beyond the number of agents and busy hour call attempts. Configurations with multiple skill groups per agent, significant call queuing, and other factors contribute to the total capacity of any individual component. Careful planning and discovery in the pre-sales process should uncover critical sizing variables, and these considerations should be applied to the final design and hardware selection.

Correct sizing and design can ensure stable deployments for large systems up to 6,000 agents and 180,000 BHCA. For smaller deployments, cost savings can be achieved with careful planning and co-resident ICM components (for example, Progger, Rogger, and Agent PG).

Additionally, designers should pay careful attention to the sizing variables that will impact sizing capacities such as skill groups per Agent. While it is often difficult to determine these variables in the pre-sales phase, it is critical to consider them during the initial design, especially when deploying co-resident PGs and Proggers. While new versions will scale far higher, the Cisco Agent Desktop Monitor Server is still limited in the number of simultaneous sessions that can be monitored by a single server when monitoring and recording are required.