Skip to main content

Overview

Prizm is designed for flexible enterprise deployment — supporting fully managed cloud, self-hosted on-premises, and hybrid configurations. The containerized architecture ensures consistent behavior across all deployment models.

Deployment Models

Cloud (SaaS)

Fully managed by DQLabs. Fastest time-to-value with automatic updates and scaling.

On-Premises

Self-hosted in your own data center or private cloud. Full control over data residency.

Hybrid

Control plane in cloud, data agents on-premises. Best of both worlds.

Architecture Principles

Prizm’s deployment architecture is built on:
  • Containerization — All components run in Docker/Kubernetes containers
  • Horizontal Scalability — Scale agent workers independently based on data volume
  • Open APIs — GraphQL and REST APIs for custom integrations and extensions
  • Secure by Default — TLS 1.3 everywhere, AES-256 at rest, zero-trust networking

Deployment Checklist

1

Choose Deployment Model

Decide between SaaS, on-premises, or hybrid based on data residency requirements and operational preferences.
2

Provision Infrastructure

Allocate compute, storage, and networking resources per the infrastructure sizing guide.
3

Configure Identity Provider

Set up SSO integration (SAML 2.0, OAuth 2.0, or LDAP) for user authentication.
4

Connect Data Sources

Add your first data connections — typically your primary data warehouse and a pipeline tool.
5

Run Initial Discovery

Execute the first asset discovery run to populate the catalog. This typically takes 15–60 minutes depending on data volume.
6

Configure Access Control

Set up user groups, roles, and domain assignments for your team.
7

Enable AI Features

Configure the AI model selection (Claude is the default) and enable Converse for your users.

Customer Onboarding

DQLabs provides a structured onboarding program for new Prizm deployments:

Onboarding Phases

PhaseDurationFocus
KickoffWeek 1Architecture review, deployment planning, credential preparation
FoundationWeeks 2–3Source connections, initial discovery, user setup
ActivationWeeks 4–5Profiling configuration, quality metric setup, alerting
ExpansionWeeks 6–8AI features enablement, governance workflows, additional sources
Steady StateOngoingReview cadence, metric tuning, new source onboarding
Contact your DQLabs Customer Success Manager for a tailored onboarding plan specific to your deployment size and data ecosystem.

Infrastructure Requirements

Minimum Requirements (Small Deployment)

ComponentSpecification
Control Plane4 vCPU, 16 GB RAM
Agent Workers2 vCPU, 8 GB RAM per worker
Database (MetaStore)PostgreSQL 14+, 100 GB SSD
Search IndexElasticsearch / OpenSearch 8+, 50 GB
Object StorageS3 / Azure Blob / GCS — 100 GB
ComponentSpecification
Control Plane8 vCPU, 32 GB RAM (HA pair)
Agent WorkersAuto-scaling pool, 2–20 workers
Database (MetaStore)Managed PostgreSQL, Multi-AZ, 500 GB
Search IndexManaged OpenSearch, 3-node cluster
Object StorageUnlimited (pay-as-you-go)
KubernetesEKS / AKS / GKE, 1.28+