HYBRID DATA PLATFORM

data lake and analytics
on one platform

Resource Lake integrates all data from the manufacturing site (Data Lake),It is an integrated data platform that simultaneously supports real-time transactions (OLTP) and deep analytics (OLAP).

RAW SOURCES
PURIFYING
STRUCTURED ASSET

Core Architecture

We provide a complete pipeline from data collection to analysis.

Data Lake

From structured data (RDB) to unstructured data (Log, Image, Sensor), it is stored in its raw form and defined and used when necessary.Schema-on-ReadProvides architecture.

  • Low-cost storage of large amounts of data
  • Accommodates all types of data (Structured/Unstructured)

OLTP

We process tens of thousands of transactions per second that occur in the field, including production performance, equipment status, and quality judgment, without delay.Data Integrity (ACID)Guaranteed.

  • Real-time data processing
  • High Concurrency Control

OLAP

By analyzing accumulated data from various dimensions, we aim to improve productivity and reduce costs.business insightsDerives . We provide high-quality datasets for AI learning.

  • Complex queries and aggregate analysis
  • Multi-dimensional data modeling

AI-Ready Data Pipeline

Beyond simple storage, LLM converts it into understandable knowledge.

STEP 01

LLM Pre-processing

Extract text from unstructured data (PDF, HWP, manual) and convert it into a structure (Markdown, JSON) that LLM can easily understand. Optimize token efficiency by removing unnecessary noise.

STEP 02

Vector Embedding

Convert the preprocessed text into a high-dimensional vector and store it in Vector DB (Milvus, Chroma). Performs an indexing process for semantic search.

STEP 03

Knowledge Service

Provides optimal context to LLM through RAG (Retrieval-Augmented Generation) pipeline. We provide knowledge to various applications in the form of API.

Quick Answers

Key questions about Resource Lake data integration

How fragmented manufacturing data becomes connected and usable.

What is Resource Lake?

It collects, cleans, and standardizes data from equipment, sensors, and legacy systems so MES, dashboards, and AI can use a shared data foundation.

Can it handle different data formats?

Fields and units from each source are mapped into a common data model so different system structures can be connected consistently.

How is the collected data used?

It supports production monitoring, quality analysis, equipment health, management metrics, and reliable data for RAG and AI analysis.

Reviewed by: IYULAB technical team
KRKO