Products

Products / SparklingSoDA 

SparklingSoDA

SparklingSoDA Concept & Vision

AI DevOps Portal For Enterprise.

SparklingSoDA is AI DevOps Portal which supports analysis and operations at once for enterprises’ optimal business decision. It helps you develop, distribute, and operate AI models and so you can transform all outputs into your assets and manage them. It provides an overall environment for AI-based analysis and operations such as auto-generation from insufficient data for analysis, self-evolving strategy recommendation functions.

SparklingSoDA : Service Architecture

All about AI DevOps

SparklingSoDA is optimized for overall AI-based analysis and operations environment from the management of accounts, packages, source versions, and data to application, distribution, operations of development outputs. 

SparklingSoDA v3.0 Introduction

SparklingSoDA is AI analysis and forecast platform for enterprises. It maximizes the efficiency of operations and management by building an analysis environment and by supporting the development, distribution, operations of AI models in a single solution. It consists of DevAI, AIOps, and DataDam(separate option).

 DevAI

Analysis model development/training based on Machine Learning.



Built-in Basic Algorithms for various AI analyses .

-Deep learning, Random Forest, GBM, etc.


Parallel processing support for high performance analytics.

Support selective operation processing of CPU and GPU when training models.

Support isolation of analysis environment to avoid conflicts between versions of Open S/W.

Internal repository configuration and mirroring service for open packages.


Individual disk allocation per analysis project, analysis code configuration management and Disk Allocation.

Transformation of Trained Models to Intellectual assets.

Built-in (RStudio Server Pro, a separate purchase required)
Analysis model development/training based on Machine Learning

  Built-in Basic Algorithms for various AI analyses.

   - Deep learning, Random Forest, GBM, etc

  Parallel processing support for high performance analytics.

  Support selective operation processing of CPU and GPU when training models.

  Support isolation of analysis environment to avoid conflicts between versions of Open S/W.

  Internal repository configuration and mirroring service for open packages.

  Individual disk allocation per analysis project, analysis code configuration management and Disk Allocation. 

  Transformation of Trained Models to Intellectual assets.

  Built-in (RStudio Server Pro, a separate purchase required).

 AIOps 

Model Distribution and Monitoring.



Real-time distribution system of analytics models to the operating environment.


Provide the system for model version management and performance comparison.
Reports to confirm the outputs of models.

Service functions without interruption to model service.


Running function by each model service process.
-Solve the bottleneck problem of CPU, GPU resources when running on a single process.

Monitoring function by model service and by user authority.

Simple management by supporting step-by-step logs for model integrated operation management.
Error alert and monitoring function.

Built-in (Shiny Server Pro, a separate purchase required)
Model Distribution and Monitoring

  Real-time distribution system of analytics models to the operating environment.

  Provide the system for model version management and performance comparison.

  Reports to confirm the outputs of models.

  Service functions without interruption to model service.

  Running function by each model service process

  - Solve the bottleneck problem of CPU, GPU resources when running on a single process.

 Monitoring function by model service and by user authority.

 Simple management by supporting step-by-step logs for model integrated operation management.

  Error alert and monitoring function.

 Built-in (Shiny Server Pro, a separate purchase required).

 DataDam

Big data storage and management



Storage of Structured, unstructured data



Support Lamda Architecture for real-time batch-based service support

Ensure optimal performance to support Scale-out/up considering data expansion


Support optimal structure for AL analysis with parallel distributed processing and storage structure

Uninterrupted service supporting redundancy for node failure
Optional installation
Big data storage and management

  Storage of Structured, unstructured data.

  Support Lamda Architecture for real-time batch-based service support.

  Ensure optimal performance to support Scale-out/up considering data expansion.

  Support optimal structure for AL analysis with parallel distributed processing and storage structure.

 Uninterrupted service supporting redundancy for node failure.

  Optional installation.

 Service Repository

Offer Built-in model references




Various built-in analysis model references for AI analysis





Two types of models.
-CPC(Customer-Product-Channel) Analysis and AI quality control - are provided and will be expanded continuously
Built-in
Offer Built-in model references

  Various built-in analysis model references for AI analysis.

  Two types of models.

       - CPC(Customer-Product-Channel) Analysis and AI quality control are provided and will be expanded continuously.

 Built-in

SparklingSoDA v3.0 Architecture

SparklingSoDA v3.0 S/W Components
Analysts work only for analytics tools, system operators only need to install and monitor.

SparklingSoDA modules are all encapsulated. The various open sources and optimization features we provide are all in the black box and in our control. Analysts only focus on analysis after installation, and system operators only need to monitor resources and others.

 SparklingSoDA Analytics environment (DevAI)

 SparklingSoDA Operating environment (AIOps)

 SparklingSoDA Big Data platform (DataDam)

SparklingSoDA Implementation Effect

Immediately jump into AI company.

After successfully implement SparklingSoDA, when executives and employees make a variety of decisions, they (so called “Self Service BI” or “Citizen analyst”) can use AI technology to realize Real-Time Enterprise and differentiate themselves, secure competitiveness, and develop new services.  It also enables you to respond to new markets such as sales of data and algorithms.

AgileSoDA


AgileSoDA supports companies to transform AI capabilities and output into internal intellectual assets and to resolve data quality problems. It also provides algorithm-based technologies and services including reinforcement learnings for optimal business decision.

Contact US