Revision as of 11:56, June 20, 2017 by Xavier (talk | contribs)
Jump to: navigation, search

Xavier:Reviews

Deploying GIR with Remote Agents - GIR Interoperability Requirements

Overview

This document outlines the interoperability requirements for Genesys Interaction Recording (GIR) to support SIP Server Remote Agents using devices or switches that are already supported by Genesys SIP Server. To support this deployment mode, the deployment must adhere to some architectural design guidelines outlined in this document.

In addition, for this deployment mode to be supported, the customer solution team must validate and document a test report for GIR interoperability to ensure that the solution is compatible with GIR. The test scenarios for validation are outlined in this document. These scenarios are focused on core GIR capabilities because Quality Management and Interaction Analytics interoperability are implicitly supported if GIR functions as expected.

Important

The scope of this document applies to deployments that contain only SIP Server and no T-Server in the deployment. Deployments with T-Server are not included in the scope of this document and are considered an unsupported configuration.

In addition, the focus of this document is the use of SIP Server for Contact Center-only deployments. GIR does not support Enterprise PBX deployments where SIP Server is used as a PBX. This means while it is technically possible for GIR to record non-agent DNs (DN with no agent login or person objects), this is not a valid deployment mode of GIR as recorded calls will not contain any information to allow other GIR-related workflows for name-based search, Quality Management, or Speech Analytics.

Deployment Model

In SIP Server deployments where a third-party switch is used, agent devices will be managed through a third-party switch. SIP Server is responsible for managing the agent login and routing calls to agents through the extension DN, while the third-party switch handles SIP registration. The extension DN in this type of model is not a SIP device directly, but rather managed by a third-party switch. From the SIP Server perspective, the third-party switch acts just like the device allowing SIP Server to bridge the call through Media Control Platform (MCP) to perform active recordings. SIP presence is not required for recording to work.

In this deployment model, all TEvent generated by SIP Server must work just like a regular SIP Server deployment. As GIR is dependent on ICON DB for collecting party events and attached data, having the full set of TEvents allows GIR to be compatible with remote agent deployment.

The following diagrams show the call paths between SIP Server, third-party switch, agent device, and MCP.

SIP Server Remote Agents with WDE SIP Server Remote Agents with WDE

SIP Server Remote Agents with WDE (recording) SIP Server Remote Agents with WDE (recording)

SIP Server Remote Agents with WWE SIP Server Remote Agents with WWE

SIP Server Remote Agents with WWE (recording) SIP Server Remote Agents with WWE (recording)

Inbound calls

Inbound calls to SIP Server arrives on a Trunk DN first . The call can be directed to a routing point DN for routing to an agent through the extension DN. To simplify configuration of the contact for extension DN, a softswitch DN can be defined so that all extension DNs can share the same configured contact information.

Outbound calls

Outbound contact solution that uses TMakePredictive call against SIP Server allows SIP Server to record outbound campaigns with agents. For manual outbound calls, agents must use third-party call control through Workspace Desktop Edition (WDE) or Workspace Web Edition (WWE) to place outbound calls. For more details, refer to Solution Limitation.

Nailed-up agents

Nailed-up agents are supported if the SIP Server supports the SIP device.

Recording controls

WDE and WWE have native support for displaying recording indications and recording controls for dynamic recording.

IVR Recording

IVR recording is supported in this deployment with Trunk Group DN as IVR (using agent recording method) or with PlayApplication treatments (with VoiceXML recording control).

Screen Recording

Screen recording is compatible for voice calls that are recorded on SIP Server. GIR monitors for any calls that are recorded by SIP Server and can trigger screen recording based on internal rules.

Hot seating support

Hot seating is supported for remote agent deployment for WDE and WWE.

Third-party Desktop support

For custom desktop applications to inter-operate with GIR screen recording, the desktop must integrate with Screen Recording Service (SRS) API as outlined in Genesys Interaction Recording API Reference.

Refer to the following sections for more details on the SRS integration:

The custom desktop must handle the following features of Screen Recording Service:

Function area Feature Description
Login Hot seating Whenever the agent changes place (either during login or after login), the customer desktop must report the DN, Place, and Switch to SRS.
Login Basic authentication Pass username and password during login to SRS.
Security Cross origin resource sharing. For web-based desktop.
Security CSRF For web-based desktop.
Security TLS Custom desktop must support Transport Layer Security (TLS) if SRS is configured to use TLS.
Failover Site failover Custom desktop must report to SRS when voice is switched over from one data center to another. A site failover usually includes a change of DN, Place, Switch, or the address of Interaction Recording Web Services (RWS). The custom desktop must report these changes to SRS.

Solution Limitations

In a SIP Server Remote Agent deployment, agents have the ability to place calls (transfers, consults, or conference) using third- party call control through WDE or WWE, or use the device directly to place calls (first-party call control).

Calls that are placed (transfers, consults, or conference) through WDE or WWE with third-party call control will be recorded by SIP Server and are accessible to GIR.

All calls that are placed through the device directly (first-party call control) will not be recorded by SIP Server and therefore are not accessible to GIR. Calls that are placed directly using 1pcc will not be supported as shown in the following figures.

Solution limitation switch.png

Solution limitation SBC MGW.png

Interoperability Requirements

The following are the GIR interoperability requirements:

  • Remote agents must have an agent login on the SIP Server switch.
  • Remote agents must have an extension DN on the SIP Server switch.
  • Softswitch DN can be used to share a common contact for extension DNs.
  • Remote agents must use third-party call control (through WDE or WWE) to place, consult, transfer, or conference all outbound calls or calls to other agents.
  • Full time recording can be controlled through extension DN, softswitch DN, or agent login.
  • Enabling recording on inbound Trunk DN is not supported.
  • Selective recording can be controlled by a routing strategy through TRouteCall.
  • Dynamic recording (start, stop, pause, or resume) can be controlled by desktops (WDE or WWE).
  • SIP Server configuration for GIR recording must be used. For configuration details, refer to Deploying SIP Server for GIR.
  • GVP configuration for GIR recording must be used. For configuration details, refer to Deploying Genesys Voice Platform for GIR.
  • ICON and ICON DB for GIR recording must be used for interoperability testing. For configuration details, refer to Deploying Interaction Concentrator for GIR. Note that the new indexes recommendation need not be used for the purpose of functional testing.
  • Third-party custom desktop applications must conform to Screen Recording Service API for handling screen recording.
  • If screen recording is implemented (with a third-party desktop), screen recording must have the ability to support hot seating.
  • IVR Recording is supported for Trunk Group DN as IVR (using agent recording method) or with PlayApplication treatments (with VoiceXML recording control).

Solution Test Plan

To validate interoperability with GIR, the complete GIR setup must be installed to validate if both voice and screen recordings are working. GIR must include the following components:

  • Recording Processor
  • Muxer
  • WebDAV for recording storage
  • Recording Web Services node
  • Cassandra for RWS
  • Elasticsearch for RWS
  • GAX recording plug-in
  • Recording Crypto Server
  • SpeechMiner components
  • Screen Recording Service on agent desktop

The following test areas are expected to be performed using WDE.

Test Case Procedure
Test area - Inbound
Inbound call to agent
  1. Configure extension DN with record=true.
  2. Place inbound call from Trunk DN to agent extension.
  3. Agent answers the call.
  4. Customer hangs up the call.
  5. Agent can see recording indicator is turned on (red) during the call.
  6. Search for call from SpeechMiner after 2 minutes and the Agent name should be displayed for the recording.
  7. Look for attached data in the event history. At a minimum you should see GSIP_RECORD and GSIP_REC_FN).
Inbound call to agent; recording controls
  1. Configure extension DN with record=true.
  2. Place inbound call from Trunk DN to agent extension.
  3. Agent answers the call.
  4. Agent pauses recording using recording controls. Agent must see the icon changed to paused.
  5. Agent resumes recording using recording controls. Agent must see the icon changed to recording.
  6. Customer hangs up the call.
  7. Search for the call from SpeechMiner after 2 minutes and the Agent name should be displayed for the recording Playback the recording; for the period where the call is paused, the audio should be silent.
Inbound call to agent; recording controls
  1. Configure extension DN with record=false.
  2. Place inbound call from Trunk DN to agent extension.
  3. Agent answers the call.
  4. Agent must see that the recording is not started at this point.
  5. Agent starts recording using recording controls. Agent must see the icon changed to recording.
  6. Customer hangs up the call.
  7. Search for the call from SpeechMiner after 2 minutes and the Agent name must be displayed for the recording.
Inbound call to routing point; route to agent with selective recording
  1. Configure extension DN with record=false.
  2. Place inbound call from Trunk DN to Routing Point DN.
  3. A routing strategy would call RouteCall with extension record=destination.
  4. Agent answers the call.
  5. Agent hangs up the call.
  6. Agent can see recording indicator is turned on (red) during the call.
  7. Search for call from SpeechMiner after 2 minutes and the Agent name should be displayed for the recording.
Inbound call to agent; single step transfer
  1. Configure extension DN with record=true for both agent1 and agent2.
  2. Place inbound call from Trunk DN to agent1 extension.
  3. Agent1 answers the call.
  4. Agent1 uses WDE to perform a single step transfer to agent2.
  5. Agent2 answers call. Agent1’s call must be released at this point.
  6. Customer hangs up the call.
  7. Search for call from SpeechMiner after 2 minutes and the Agent name should be displayed for the recording.
    Two media files must be presented for this recording; each segment should be the recording for each agent.
Inbound call to agent; single step conference
  1. Configure extension DN with record=true for both agent1 and agent2.
  2. Place inbound call from Trunk DN to agent1 extension.
  3. Agent1 answers the call.
  4. Agent1 uses WDE to perform single step conference to agent2.
  5. Agent2 answers the call.
  6. Agent1 hangs up the call.
  7. Customer hangs up the call.
  8. Search for call from SpeechMiner after 2 minutes and the Agent name should be displayed for the recording.
    Two media files should be presented for this recording. The first segment with agent1 must contain audio during the conference. The second segment with agent2 must contain audio after agent1 releases the call.
Inbound call to agent; 2-step transfer
  1. Make sure SIP Server configuration record-consult-calls is set to true.
  2. Configure extension DN with record=true for both agent1 and agent2.
  3. Place inbound call from Trunk DN to agent1 extension.
  4. Agent1 answers the call.
  5. Agent1 uses WDE to perform consultation to agent2.
  6. Agent2 answers the call.
  7. Agent1 completes the transfer.
  8. Customer must be talking to agent2.
  9. Customer hangs up the call.
  10. Search for call from SpeechMiner after 2 minutes and there should be two separate entries for the call.
    The first entry would contain 2 media files: Customer > agent1; Customer > agent2. The second entry is the consultation recording.
Inbound call to agent; 2-step conference
  1. Make sure SIP Server configuration record-consult-calls is set to true.
  2. Configure extension DN with record=true for both agent1 and agent2.
  3. Place inbound call from Trunk DN to agent1 extension.
  4. Agent1 answers the call.
  5. Agent1 uses WDE to perform consultation to agent2.
  6. Agent2 answers the call.
  7. Agent1 completes the conference.
  8. Customer must be talking to agent1 and agent2.
  9. Agent 1 hangs up the call. Customer must be talking to agent 2 only.
  10. Customer hangs up the call.
  11. Search for call from SpeechMiner after 2 minutes and there should be two separate entries for the call.
    The first entry would contain 2 media files: Customer > agent1 + conference; Customer > agent2. The second entry is the consultation recording.
Test area – Outbound
Agent uses WDE to place 3rd party call control (3pcc) call to another agent
  1. Configure extension DN for Agent1 with record=true.
  2. Agent1 uses WDE to place an outbound call to another agent2 (this is a 3pcc call).
  3. Agent2 answers the call.
  4. Agent2 hangs up the call.
  5. Both agents can see recording indicator is turned on (red) during the call.
  6. Search for call from SpeechMiner after 2 minutes and the agent1 name should be displayed for the recording.
Agent uses WDE to place 3rd party call control call to outside number
  1. Configure extension DN for Agent1 with record=true.
  2. Agent uses WDE to place an outbound call to an outside number (this is a 3pcc call).
  3. Customer answers the call.
  4. Customer hangs up the call.
  5. Agent can see recording indicator is turned on (red) during the call.
  6. Search for call from SpeechMiner after 2 minutes and the agent name should be displayed for the recording.
Test area - screen recording

Before proceeding with the tests, ensure that Screen Recording Service is installed on agent desktops and enable screen recording.

Inbound call to agent
  1. Configure extension DN with record=true.
  2. Place inbound call from Trunk DN to agent extension.
  3. Agent answers the call.
  4. Customer hangs up the call.
  5. Agent can see recording indicator is turned on (red) during the call.
  6. Search for call from SpeechMiner after 2 minutes and the Agent name should be displayed for the recording.
  7. Look for attached data in the event history. At a minimum, you should see GSIP_RECORD and GSIP_REC_FN.
  8. Wait for 5 minutes for muxer to complete muxing process and access the same recording on SpeechMiner. The screen recording should now be accessible.
Inbound call to agent; recording controls
  1. Configure extension DN with record=true.
  2. Place inbound call from Trunk DN to agent extension.
  3. Agent answers the call.
  4. Agent pauses recording using recording controls. Agent must see the icon changed to pause.
  5. Agent resumes recording using recording controls. Agent must see the icon changed to recording.
  6. Customer hangs up the call.
  7. Search for the call from SpeechMiner after 2 minutes and the Agent name should be displayed for the recording.
  8. Playback the recording. Note that for the period when the call is paused the audio should be silent.
  9. Wait for 5 minutes for muxer to complete muxing process and access the same recording on SpeechMiner. The screen recording should now be accessible. The paused periods should be blacked out.
Inbound call to agent; single step transfer
  1. Configure extension DN with record=true for both agent1 and agent2.
  2. Place inbound call from Trunk DN to agent1 extension.
  3. Agent1 answers the call.
  4. Agent1 uses WDE to perform single step transfer to agent2.
  5. Agent2 answers call; agent1’s call must be released at this point.
  6. Customer hangs up the call.
  7. Search for call from SpeechMiner after 2 minutes and the Agent name should be displayed for the recording two media files should be presented for this recording; each segment should be the recording for each agent.
  8. Wait for 5 minutes for muxer to complete muxing process and access the same recording on SpeechMiner. The screen recording should now be accessible for both segments.
Inbound call to agent; single step conference
  1. Configure extension DN with record=true for both agent1 and agent2.
  2. Place inbound call from Trunk DN to agent1 extension.
  3. Agent1 answers the call.
  4. Agent1 uses WDE to perform single step conference to agent2.
  5. Agent2 answers the call.
  6. Agent1 hangs up the call.
  7. Customer hangs up the call.
  8. Search for call from SpeechMiner after 2 minutes and the Agent name should be displayed for the recording two media files should be presented for this recording.
    The first segment with agent1 should contain audio during the conference. The second segment with agent2 should contain audio after agent1 releases the call.
  9. Wait for 5 minutes for muxer to complete muxing process and access the same recording on SpeechMiner. The screen recording should now be accessible for both segments.
Inbound call to agent; 2-step transfer
  1. Make sure SIP Server configuration record-consult-calls is set to true.
  2. Configure extension DN with record=true for both agent1 and agent2.
  3. Place inbound call from Trunk DN to agent1 extension.
  4. Agent1 answers the call.
  5. Agent1 uses WDE to perform 2 step transfer to agent2.
  6. Agent2 answers the call.
  7. Agent1 completes the transfer.
  8. Customer must be talking to Agent2.
  9. Customer hangs up the call.
  10. Search for call from SpeechMiner after 2 minutes and there should be two separate entries for the call.
    The first entry would contain two media files: Customer > agent1; Customer > agent2. The second entry is the consultation recording.
  11. Wait for 5 minutes for muxer to complete muxing process and access the same recording on SpeechMiner. The screen recording should now be accessible for both segments for the first entry and a screen recording in the second entry.
Inbound call to agent; 2-step conference
  1. Make sure SIP Server configuration record-consult-calls is set to true.
  2. Configure extension DN with record=true for both agent1 and agent2.
  3. Place inbound call from Trunk DN to agent1 extension.
  4. Agent1 answers the call.
  5. Agent1 uses WDE to perform consultation to agent2.
  6. Agent2 answers the call.
  7. Agent1 completes the conference.
  8. Customer must be talking to agent1 and agent2.
  9. Agent 1 hangs up the call. Customer must be talking to agent2 only.
  10. Customer hangs up the call.
  11. Search for call from SpeechMiner after 2 minutes and there should be two separate entries for the call.
    The first entry would contain two media files: Customer > agent1 + conference; Customer > agent2. The second entry is the consultation recording.
  12. Wait for 5 minutes for muxer to complete muxing process and access the same recording on SpeechMiner. The screen recording should now be accessible for both segments for the first entry and a screen recording in the second entry.
Comments or questions about this documentation? Contact us for support!