Skip to main content

AWS BYOC Deployment

This guide describes how the Viven Digital Twin platform is deployed within the customer's own Virtual Private Cloud (VPC) on Amazon AWS. This deployment model is also known as Bring Your Own Cloud (BYOC).

Advantages of this setup:

  • Customer data resides within their own cloud and never leaves it
  • Application logs are stored within the customer's cloud
  • Customer has full control over the deployment

Customers provide a new, blank AWS account where Viven deploys the complete Digital Twin stack. All data, including application logs, is stored in the customer's cloud. A limited amount of diagnostic data is pushed to Viven's deployment command center for monitoring purposes.

Need help?

At any point, reach out to your Viven representative or support@viven.ai.


Overview

StepTask
Step 1Create a dedicated AWS sub-account
Step 2Share account access details with Viven
Step 3Set up ACM certificates
After setupPost-infrastructure steps

Step 1: Create a Sub-Account

Create a new sub-account under your AWS Organization to be used exclusively for the Digital Twin deployment.

  1. Log in to your AWS Management Account (the parent/organization account)
  2. Open AWS Organizations in the console
  3. Select Accounts → Add an AWS Account
  4. Choose Create an AWS Account (not Invite) and configure:
    • Account Name: DigitalTwin-Prod (or your naming convention)
    • Email: A distribution or admin email you control (e.g. aws-admin@yourdomain.com)
  5. Complete creation and wait for the account to appear in your accounts list
note

No billing or charges will occur at this stage. Services will only be activated later during Viven's deployment.


Step 2: Share Access Details with Viven

Once the account is created, provide your Viven representative with the following information:

DetailExample
Preferred domain and company nameacmeinc.com, Acme Inc
AWS Account ID (12 digits)123456789012
AWS Regionus-west-2
IAM user with admin accessprod-dt
ARN of the IAM rolearn:aws:iam::123456789012:role/dt-admin-access-role

With this information, Viven can begin infrastructure deployment and typically completes it within a few business days.

Create the prod-dt IAM User

  1. Go to IAM → Users → Create user
  2. Enter the username prod-dt and click Next
  3. Under Permissions options, select Attach policies directly, then select AdministratorAccess and click Next
  4. Click Create User

Create the dt-admin-access-role IAM Role

Run the CloudFormation template provided by Viven. This role will be used by Viven engineers to provision the Digital Twin stack in your account, and thereafter only by Viven's on-call team to resolve critical issues.


Step 3: Set Up ACM Certificates

Two ACM certificates are required:

  • One in your deployment region (for the Application Load Balancer)
  • One in us-east-1 (for CloudFront distributions)

For each certificate, set the fully qualified domain name to *.<your_domain>.

Once created, share both certificate ARNs and your domain name with your Viven representative.


Steps After Infrastructure Setup

Once your Digital Twin infrastructure is live, your Viven representative will follow up on:

  1. Connector setup — Connect Digital Twin to your enterprise data sources (Google Workspace, Microsoft 365, Salesforce/Slack, Atlassian, and more)
  2. SSO setup — Provide your Identity Provider (IdP) configuration: entity ID, SSO URL endpoints, and X.509 certificates
  3. LLM setup — Digital Twin uses OpenAI models. OpenAI does not train on inputs or outputs from business customers
  4. User limiting — Access is best restricted via SSO. The Admin Console also allows admins to limit users directly

Appendix

AWS Services

AWS ServicePurpose
Amazon CloudFrontSecurely distributes the application with low latency
AWS Elastic Kubernetes Service (EKS)Hosts the web application and processor workers
Amazon EC2Compute instances and Airflow
Amazon RedshiftStores application and access logs
Amazon RDS (Aurora)Stores user information and application configuration
Amazon Route 53Manages domain records
Elastic Load BalancingBalances traffic across EC2 instances
Amazon ElastiCacheCaches results of expensive I/O calls and computations
Amazon DynamoDBPrimary datastore for processed data
Amazon ECRStores container images for deployment
Amazon S3Stores raw ingested documents, builds, and artifacts
AWS IAMCreates and manages roles used by the platform
AWS KMSCreates and manages encryption keys
Amazon SESSends emails from the application
Amazon SQSQueues tasks for background workers
Amazon SNSPub/sub notifications to multiple consumers
Amazon KinesisStreams application logs to Redshift
AWS Secrets ManagerStores application secrets (API tokens, etc.)
Amazon CloudWatchCollects metrics and logs for monitoring
Amazon OpenSearch ServiceIndexes documents and supports lexical and vector search
PostgreSQL (RDS)Stores Airflow logs

Document Ingestion Architecture

1. Scheduled Orchestration

An Airflow DAG runs on a periodic basis and queues an ingestion job for the EKS Worker cluster for each connected source system (Gmail, Calendar, Google Drive, etc.).

2. EKS Worker Cluster Processing

StageDescription
Document IngestWorker calls external source systems via Connectors and pulls documents; raw files are stored in S3
Document StoreDocument metadata is stored in DynamoDB
Document ProcessDocuments are enriched via the LLM service; enriched data is stored in DynamoDB
Document IndexDocuments and their embeddings are indexed in Amazon OpenSearch for search and retrieval

Application Logs and Configuration Access

Application logs are stored in Redshift and application configuration is stored in RDS — both within the customer's account. Viven's tooling provides easy interfaces for accessing this data when troubleshooting issues or enabling new features. A dedicated user login is provisioned in the deployed application for this purpose.