Netcell Event Driven Platform
Want to market your business and serve your customers better?
Let's talk:
NETCELL-EDA
NETCELL‑EDA is a multi‑tenant, event‑driven AI platform designed to ingest operational events, normalize them, process them through microservices, generate enriched profiles, and deliver real‑time inference.
The system is built on Kafka, Kubernetes, and a dual‑layer feature store (online + offline), enabling scalable, low‑latency AI workloads.
The architecture is modular, resilient, and optimized for incremental rollout — starting with an MVP slice and expanding into a full multi‑tenant deployment. .
Netcell EDA ArchitectureNETCELL-EDA Key Components
Core architecture Layers:
Integration Layer
Events Hub (Kafka + routing)
Event Processing Layer (microservices)
Profiling & AI Layer
Feature Store (online + offline)
Model Registry
Analytics & Dashboards
Platform Services (Config, Identity, Observability)
NETCELL-EDA Configuration
Integration Layer (Edge Services)
These services sit closest to external systems.
Integration Pull Service
Polls external systems by interval
Converts raw data into canonical events
Ideal for legacy systems
Integration Agent Service
Runs on client application servers
Hooks into DAL events
Pushes events directly to Events Hub
Integration API Gateway
External systems push events via REST
Handles auth, throttling, validation
Output: Canonical events → sent to Events Hub.
Read moreNETCELL-EDA Design and Modeling
Model registry design
Goal:
Central place to store, version, and manage all AI models (risk, churn, engagement, etc.) so Inference and Profiling Services can use them safely.
Core concepts
Model:
a single trained artifact (file).
Model Type:
e.g. Risk, Churn, Engagement.
Version:
e.g. risk_v3, churn_v2.
Status:
draft, validated, production, deprecated.
Tenant scope:
global or per TenantId.
Storage
Metadata table (SQL / Hive):
Model files stored in:
S3 / HDFS / Cloudera storage.
And API , Usage...
Read more
NETCELL-EDA AI Model Governance Framework
Purpose
This policy defines how machine learning models are developed, validated, deployed, monitored, and retired within Netcell‑EDA. It ensures consistency, safety, transparency, and compliance across all tenants and environments.
Scope
This policy applies to:
All ML models used for profiling, scoring, inference, or analytics
All environments (dev, staging, production)
All tenants
All data science and engineering teams
NETCELL-EDA Security Architecture
Core principles
Least privilege:
All services, users, and tenants get only the permissions they need.
Defense in depth:
Network, identity, data, and application layers all enforce controls.
Tenant isolation:
Strong logical isolation via topics, namespaces, and storage boundaries.
Auditability:
Every sensitive action is logged and traceable.
NETCELL-EDA Business Continuity & Recovery Strategy
Purpose This plan ensures Netcell‑EDA can recover from catastrophic failures while maintaining data integrity, tenant isolation, and service continuity.
Recovery Objectives
RPO (Recovery Point Objective):
Maximum data loss: ≤ 5 minutes (Kafka replication + DB snapshots)
RTO (Recovery Time Objective):
Maximum downtime: ≤ 30 minutes (automated failover)
NETCELL-EDA Services Architecture
Purpose: Features for entities (Contact, Account, etc.), run the right model, return scores fast enough to be used in event processing.
Used by:
Contact Events Processor
Account/Billing/Security processors
Analytics services (on-demand scoring)
NETCELL-EDA Events Processor
This microservice design is perfect for Cloudera
Kafka integration is native
Schema Registry fits perfectly
Ranger can enforce tenant isolation
Spark/Flink can generate offline features
Iceberg tables store historical data
Kubernetes handles scaling
AI pipeline plugs into Cloudera storage
NETCELL-EDA Feature Store Architecture (Enterprise Ready)
What the Feature Store actually does
It solves three problems:
Consistency
Training and inference use the same features.
Reusability
Multiple models (risk, churn, engagement) can use the same features.
Scalability Features are stored once, reused many times.
NETCELL-EDA Microservices Map (Strategy)
Kafka topic strategy
Define core
Domains, Contacts, Accounts, Billing, Security
Name topics per domain
contacts-events, accounts-events, billing-events, security-events
Partitioning strategy
Start with 4–8 partitions per topic.
Scale up as traffic grows.
Tenant isolation model:
Per-tenant topics,
Shared topics with TenantId in message,
Governance,
Event schema registry,
Define canonical event format
All events share a base structure:
Document DB profile structure,
AI profiling pipeline
NETCELL-EDA Deployment Map
Microservices (Kubernetes Pods)
A. Contact Events Processor
Consumes:
contacts-events topic
Group:
contact-processor-group
Replicas:
4 pods
Responsibility:
Feature building
Feature store updates
Inference call
Profile update logic
Profiling Service,
Replicas: 3 pods
Responsibility:
Feature building,
Feature store updates,
Inference call,
Profile update logic
NETCELL-EDA Cloudera Integration
Cloudera Integration Blueprint (Enterprise Architecture)
This is how entire platform plugs into Cloudera’s ecosystem.
This blueprint is designed for large organizations, multi‑tenant, high‑scale, secure.
NETCELL-EDA Business Automation Platform
Business Automation Powered by Netcell‑EDA
Continuous, intelligent, event‑driven engagement for every business
Modern businesses lose time, customers, and opportunities simply because they cannot respond fast enough.
Netcell‑EDA changes that.
By combining real‑time event processing with automated decisioning, it enables a new generation of marketing automation that works alongside your team, increases daily productivity, and ensures no customer or operational moment is ever missed.
NETCELL-EDA Task Management Platform
Netcell Task Management System A unified, AI‑powered workspace that automates task creation, routes work intelligently,
enforces SLAs, and gives every team a modern interface to execute with clarity and speed.
Built on Scope*,
it delivers a fast, fluid, ultra‑modern experience backed by a robust microservices architecture and real‑time KPI insights.
Netcell Task Management System Transforms every customer event into actionable work using Netcell EDA and domain‑specific AI Agents.
It unifies marketing automation, sales operations, service workflows, and project execution into one intelligent workspace — powered by Scope* for a fast, fluid, modern experience.
Let's talk and examine how the automated system can contribute to your business.
Send a WhatsApp message, call or leave details:
About Netcell
Netcell helps businesses turn every customer interaction, operational trigger, and marketing event into intelligent, actionable work.
We believe companies don’t fail because they lack data — they fail because they lack execution.
Powered by Netcell EDA, domain specific AI Agents, and a unified Scope SPA workspace, Netcell transforms raw signals into tasks that route themselves to the right team with the right priority and SLA.
Netcell is the creator of Event Driven Business Execution (EDBE) — a new category designed to help organizations operate faster, smarter, and more consistently.
Our mission is simple: Make business execution intelligent.
Our Platform
Netcell integrates advanced event processing (Netcell EDA), domain‑specific AI Agents, intelligent routing, SLA enforcement, and a unified Scope SPA workspace.
This architecture enables organizations to:
- Improve operational efficiency
- Strengthen customer engagement
- Reduce execution gaps
- Enhance cross‑department collaboration
- Maintain real‑time visibility into performance and outcomes
Contact
We are at your disposal with a smile in any of the following ways:
Office address:
Stock Exchange Complex: 33 Jabotinsky St.
Email:
s-support@my-t.co.il
Phone:
074-70-20-802