Amazon Connect Supported Architectures


Overview

CX Intelligence for Amazon Connect is currently offered in 2 supported architectures:

1. Single Tier - eMite Cloud

  • Can be purchased from the AWS Marketplace here


2. Single Tier - Customer Cloud

  • Deployed in the customer’s AWS cloud using an eMite provided AWS CloudFormation (template)

  • eMite’s CloudFormation template can be configured to:

    • Use an existing VPC and subnets

    • Customize Security Group Settings

    • Use existing Kinesis streams connected to your Amazon Connect solution

    • Deploy an Application Load Balancer

  • Single Sign On (SSO) is also supported with various SAML 2.0 vendors listed in SAML Single Sign On (v7).

The following architecture/deployment options are currently not supported as standard:

  • Multi-Tier (Highly Available) architectures

  • Horizontally scalable Auto-scaling architectures

  • Manual assisted deployments (where eMite deploys the solution manually via the customer’s AWS console)

  • Terraform Deployment Scripts

  • Integration with AWS Lambda / S3 Buckets to process CTR Records

  • Connecting two or more Amazon Connect instances to a single eMite instance

  • Configuring Resource Policies to the AWS Elasticsearch Domain


The following diagrams depict the 2 supported architectures at a high level:

Single Tier - eMite Cloud

 

Single Tier - Customer Cloud


Data Connection Overview

eMite utilizes 2 adapters to onboard reporting data from Amazon Connect:


1. Amazon Connect Adapter

  • Connects to the Amazon Connect instance directly

  • Gets Queue Observations and Queue Metrics data


2. Amazon Kinesis Adapter

  • Connects to the configured Amazon Kinesis Streams

  • Gets real-time AgentEvents data

  • Can store this on ElasticSearch indices (optional)

  • Gets Contact Trace Record data

  • Stores this data on ElasticSearch


Frequently Asked Questions


1. Why doesn't eMite use Kinesis Firehose?

Previously, we used Kinesis Firehose to take CTR data from Kinesis stream and send it to Elasticsearch.
Now, we use the eMite adapters to get the CTR data from Kinesis stream and send it to Elasticsearch.
This allows us to better process the data before storing it on Elasticsearch.


2. In the eMite hosted scenario, are there any data-out costs from the customer’s AWS instance to the eMite AWS instance?

  • Amazon Connect - No additional costs; There is no charge for API calls to the Connect instance

  • Kinesis Streams - No additional costs; GET requests to the stream are not charged


3. Are there any data-out costs from eMite AWS instance?

Customers will not be charged by AWS for this.


4. Who will configure the Kinesis streams?

The AgentEvent and Kinesis streams would be created automatically by the cloudformation template. After these are created, the customer is required to enable Data Streaming on their Amazon Connect instance and point it to the created Kinesis streams. Instructions on how to do this are provided by eMite.