Adobe Experience Platform Architecture: Complete Technical Guide

Adobe Experience Platform (AEP) is the foundation of Adobe's Experience Cloud, providing unified data management, real-time customer profiles, and AI-driven insights. Understanding its architecture is essential for successful implementations and maximizing the platform's capabilities.

AEP Architecture Overview

AEP is built on a multi-layer architecture designed for scalability, real-time processing, and enterprise-grade security:

┌─────────────────────────────────────────────────────────────────────────┐
│                        APPLICATION LAYER                                 │
│  ┌──────────┐ ┌──────────┐ ┌──────────┐ ┌──────────┐ ┌──────────┐      │
│  │ Real-Time│ │ Customer │ │ Journey  │ │ Offer    │ │ Adobe    │      │
│  │   CDP    │ │ Journey  │ │ Optimizer│ │Decisioning│ │Analytics │      │
│  └──────────┘ └──────────┘ └──────────┘ └──────────┘ └──────────┘      │
├─────────────────────────────────────────────────────────────────────────┤
│                        SERVICES LAYER                                    │
│  ┌──────────┐ ┌──────────┐ ┌──────────┐ ┌──────────┐ ┌──────────┐      │
│  │Segmentation│ │Destinations│ │ Query   │ │ Data    │ │   AI/ML  │      │
│  │ Service  │ │  Service │ │ Service │ │ Science │ │ Services │      │
│  └──────────┘ └──────────┘ └──────────┘ └──────────┘ └──────────┘      │
├─────────────────────────────────────────────────────────────────────────┤
│                         DATA LAYER                                       │
│  ┌──────────────────────────────────────────────────────────────┐      │
│  │              Unified Profile Service                          │      │
│  │  ┌─────────┐ ┌─────────┐ ┌─────────┐ ┌─────────┐             │      │
│  │  │ Profile │ │ Events  │ │ Segments│ │ Computed│             │      │
│  │  │ Attrs   │ │         │ │         │ │ Attrs   │             │      │
│  │  └─────────┘ └─────────┘ └─────────┘ └─────────┘             │      │
│  └──────────────────────────────────────────────────────────────┘      │
│  ┌──────────────────────────────────────────────────────────────┐      │
│  │              Identity Service                                 │      │
│  │  ┌─────────┐ ┌─────────┐ ┌─────────┐                         │      │
│  │  │ Identity│ │ Identity│ │ Private │                         │      │
│  │  │  Graph  │ │Namespaces││  Graph  │                         │      │
│  │  └─────────┘ └─────────┘ └─────────┘                         │      │
│  └──────────────────────────────────────────────────────────────┘      │
│  ┌──────────────────────────────────────────────────────────────┐      │
│  │              Data Lake (XDM-based)                            │      │
│  │  ┌─────────┐ ┌─────────┐ ┌─────────┐ ┌─────────┐             │      │
│  │  │ Datasets│ │ Schemas │ │ Batches │ │ Catalog │             │      │
│  │  └─────────┘ └─────────┘ └─────────┘ └─────────┘             │      │
│  └──────────────────────────────────────────────────────────────┘      │
├─────────────────────────────────────────────────────────────────────────┤
│                       INGESTION LAYER                                    │
│  ┌──────────┐ ┌──────────┐ ┌──────────┐ ┌──────────┐ ┌──────────┐      │
│  │ Web SDK  │ │Mobile SDK│ │Server API│ │  Source  │ │  Batch   │      │
│  │          │ │          │ │          │ │Connectors│ │  Upload  │      │
│  └──────────┘ └──────────┘ └──────────┘ └──────────┘ └──────────┘      │
└─────────────────────────────────────────────────────────────────────────┘

Data Ingestion Layer

The ingestion layer handles data coming into AEP through multiple channels:

Streaming Ingestion

Web SDK (Alloy.js) — Client-side JavaScript for web data collection
Mobile SDK — iOS and Android libraries
Server API — Server-to-server HTTP streaming
Edge Network — Global CDN for low-latency collection

Batch Ingestion

Source Connectors — Pre-built integrations (Salesforce, S3, Azure, etc.)
Direct Upload — CSV/Parquet file uploads via UI
API Batch — Programmatic batch uploads

                Ingestion Guardrails
                Streaming: Up to 20,000 events/second per sandbox
Batch: Files up to 100GB per batch
API calls: Rate limited per organization

            

XDM Schema Architecture

Experience Data Model (XDM) is the standardized data format for all data in AEP:

Schema Components

Classes — Define the base type (Profile, ExperienceEvent, Custom)
Field Groups — Reusable sets of fields
Data Types — Complex field definitions
Individual Fields — Scalar values with types

Schema Types

Schema Class	Use Case	Profile Enabled?
XDM Individual Profile	Customer attributes, preferences	Yes
XDM ExperienceEvent	Time-series events, interactions	Yes (linked)
XDM Business Account	B2B account data	Yes (B2B)
Custom Classes	Lookup tables, reference data	Optional

Identity Service Architecture

The Identity Service stitches together customer data across touchpoints:

Identity Namespaces

Standard Namespaces — Email, Phone, ECID, GAID, IDFA
Custom Namespaces — CRM ID, Loyalty ID, etc.
Cross-device — Links mobile/web/offline identities

Identity Graph

The identity graph connects all identifiers belonging to a single person:

                    Identity Graph Example
                    ─────────────────────
                           
                    ┌─────────────┐
                    │   Person    │
                    │   (Node)    │
                    └──────┬──────┘
           ┌───────────────┼───────────────┐
           │               │               │
    ┌──────┴──────┐ ┌──────┴──────┐ ┌──────┴──────┐
    │    ECID     │ │    Email    │ │   CRM ID    │
    │ abc123...   │ │[email protected] │ │  CUST-001   │
    └─────────────┘ └─────────────┘ └─────────────┘

⚠️ Identity Graph Limits

Each identity graph is limited to 50 identities per person. Exceeding this can cause graph collapse. See our Identity Graph Collapse Fix guide.

Unified Profile Service

The Real-Time Customer Profile merges all data into a single view:

Profile Components

Profile Attributes — Latest known values (name, email, preferences)
Experience Events — Time-series behavioral data
Segment Memberships — Audiences the profile belongs to
Computed Attributes — Derived values (lifetime value, propensity scores)

Profile Merge Policies

Merge policies determine how data is combined when identity graphs have multiple sources:

Dataset Precedence — Priority order for conflicting attributes
Timestamp Ordered — Most recent value wins
Private Graph vs Shared — Identity linking behavior

Segmentation Architecture

Segmentation Types

Type	Evaluation	Latency	Use Case
Batch Segmentation	Daily scheduled	24 hours	Large audiences, complex queries
Streaming Segmentation	Near real-time	Minutes	Simple criteria, profile changes
Edge Segmentation	Real-time at edge	Milliseconds	Same-page personalization

Destinations & Activation

Destinations allow you to activate audiences to external platforms:

Streaming Destinations — Real-time activation (Google, Meta, etc.)
File-based Destinations — Batch exports (S3, SFTP)
Personalization Destinations — Adobe Target, custom personalization

For detailed information, see our AEP Destinations Guide.

Related AEP Guides

Need AEP Implementation Help?

WegWorks specializes in Adobe Experience Platform implementations, schema design, and data architecture. Contact us for expert guidance.