Collection of various data
Connect PLC · SCADA · MES · LIMS with standard protocols and CDC
Data Lakehouse
Connect fragmented structured and unstructured data into a unified analytics foundation. Collect, refine, and store business-system, document, and equipment data, transforming it into data assets ready for analytics and AI.
OVERVIEW
Rather than just collecting data, Analytics and AI are ready for immediate use.
A data lakehouse combines the flexibility of a data lake that preserves the original, It combines the advantages of a data warehouse that reliably analyzes refined data into a single structure.
By collecting and standardizing distributed data, everything from BI analysis to AI model learning and natural language query is executed on a single data foundation.
original preservation
Reliably store data while maintaining format and provenance
Refining/Standardization
Processed to suit quality and purpose of use in Bronze–Silver–Gold stages
Analysis and AI utilization
Expansion to BI, SQL analysis, model learning, and natural language data querying
scattered dataOne analytics asset you can trustSwitch to .
CHALLENGES
Although data is accumulating, There is still no connection between the systems.
As data is separated by business system and collection and purification methods change, more time is required to prepare for analysis and AI use.
Data is distributed across business systems.
The storage structure of each system, such as ERP, MES, LIMS, and document management, is different, making it difficult to understand company data at once.
Structured data and unstructured data are managed separately.
Numerical data and documents, images, and files in the DB are separated from each other, making integrated analysis and AI learning data configuration difficult.
It is difficult to collect data without burdening operational systems.
Bulk queries and repetitive extractions can affect operational DB performance, limiting real-time data utilization.
Manual extraction and Excel collection are repeated.
As each department downloads and combines the necessary data directly, analysis preparation time increases and the possibility of errors increases.
It is difficult to trust data quality and consistency.
It is difficult to track where data comes from and how it has changed, making it difficult to use it directly in decision-making and AI models.
SOLUTION
We connect the entire data lifecycle, from collection to quality and genealogy.
Load data while reducing the burden on the operating system, and create a structure suitable for analysis and AI utilization through open standards and step-by-step refinement.
01
No-load data collection/loading
It connects structured and unstructured data with CDC, API, file, and facility protocols, minimizes the performance impact of the operating system, and synchronizes in real time and batch.
02
Bronze – Silver – Gold Tablets
Separate the purpose of each step with original preservation, purification/standardization, and integrated data for analysis.
03
Open table format
Based on Apache Iceberg, it provides an analysis data structure that is not dependent on a specific vendor.
04
Distributed SQL integrated analysis
Connect to Trino, Spark, etc. to analyze multiple data sources with standard SQL.
05
Automatic determination of data quality
Automatically checks for omissions, duplications, range errors, and consistency issues and provides notifications.
06
Lineage/Governance Tracking
Track data sources and change history and manage access rights and audit standards.
ARCHITECTURE
From source systems to analysis and AI, uninterrupted data service flow
AkashiQ integrates data collection, purification, storage, quality, and lineage, and connects existing BI and AI services to utilize the same data assets.
Enterprise Data Lakehouse Flow
Structured and unstructured data are integrated into one open data base and used for analysis, AI, and natural language queries depending on the purpose.
ERP · MES · LIMS
CDC-based operating system no-load interconnection
Document management · EDMS
API · PDF · Word · Excel · Image
Equipment · PLC / SCADA
OPC-UA based real-time data collection
Files/external data
Batch · Object · Log · Other data
AkashiQ · DATA LAKEHOUSE
Collection · Purification · Storage · Quality · Genealogy
We prepare it step by step into the form required for analysis and AI while preserving the original data.
Bronze
original preservation
Silver
Refining/Standardization
Gold
Integrations for analytics
Apache Iceberg · Open Table FormatAutomatic determination and notification of data quality (DQ)Data Lineage · Authority · Governance
BI · Report
Linkage with existing analysis tools based on standard SQL
KosmosAI
Analysis · Model learning · Prediction · Anomaly detection
SAMANDA
Natural language data query Text-to-SQL
Based platform: ORKESTRIX — Installation and operation of on-premises Kubernetes environment
AkashiQ
Collection · Purification · Storage · Quality · Lineage · Governance
ORKESTRIX
Foundation for installation and operation of data platform
KosmosAI
Analysis based on loaded data and utilization of AI models
SAMANDA
Natural language data queries and Text-to-SQL extensions
BUSINESS VALUE
Reduce the time it takes to find and combine data, and accelerate the use of analytics and AI.
01 · SINGLE ACCESS
Single access window for enterprise data
Find and utilize data scattered across systems and departments through one catalog and query path.
02 · UNIFIED ANALYTICS
Integrated analysis between systems
ERP, MES, LIMS, and document data are connected to standard SQL and analyzed together across system boundaries.
03 · ZERO IMPACT
Operating system no-load synchronization
Utilizes CDC and streaming to synchronize data in real time and batch while minimizing performance impact on operational DB.
04 · OPEN STANDARD
Linkage with existing BI and standards
Connects existing BI and analysis tools based on open table format and standard SQL.
05 · TRUSTED DATA
Trusted Data Foundation
Through data quality judgment and genealogy, we confirm the source, processing process, and change history, and increase the reliability of analysis and AI.
APPLICATIONS
여러 시스템에 분산된 데이터를 하나의 분석 · AI 기반으로 통합합니다.
업무 · 문서 · 현장 데이터를 한곳에 연결해 분석과 AI가 활용할 수 있는 공통 데이터 기반을 구축합니다.
제약 · 제조 데이터 기반
MES · LIMS · 설비 · 품질 데이터를 연결해 공정 분석과 AI 활용을 위한 기반을 구축합니다.
- 생산 · 설비 · 품질 데이터 수집
- LOT · 배치 단위 통합 분석
- 문서 · DB 데이터 통합 저장
- 분석 · AI 모델용 데이터 제공