On-premise refers to a method in which an organisation builds and operates its IT infrastructure directly in its own data centre or on its own servers, instead of using an external cloud. Adopting an on-premise model allows organisations to directly control data security and their systems.
On-premise refers to a method in which an organisation builds and operates its IT infrastructure directly in its own data centre or on its own servers, instead of using an external cloud. Adopting an on-premise model allows organisations to directly control data security and their systems.
This document technically explains the key benefits expected from adopting Clevi On-Premise and the secure isolated computing environment for the Clevi-X MDLM(Massive Multimodal Language Model) model.
1. Benefits of adopting On-Premise
Maximising data independence and security
All processes, from data input through analysis, service delivery and backup, are operated independently within the organisation’s closed internal network. → Minimises the risk of external data leakage and is optimised for data sovereignty and personal information protection
Sensitive data, such as Fine Dataset and private datasets, is isolated in a separate secure area → This facilitates data access control and auditing.
GPU/CPU servers are configured as a cluster in the analysis and computing domain → Optimised for large-scale AI computation, real-time data processing and parallel analysis
An HA(High Availability) structure is applied → Enables stable operation without service interruptions even when failures occur.
Flexible data management and scalability
The integrated database cluster and cold backup storage are separated → Enables efficient management of real-time data and long-term retention data
The analysis, service and data management domains are modularised, making expansion and maintenance easy → Enables a rapid response to business changes, data growth and the introduction of new services.
SVCE(Clevi Safe Virtural Compute Environment)
Enhanced security → Fully isolates each task and user session to prevent the risks of malicious code execution, data leakage and system compromise.
Flexibility and scalability → Dynamically adjusts the number of instances and resource allocations as needed, providing customised environments for various AI workloads (text, images, code, etc.).
Operational efficiency → Maximises infrastructure utilisation through automated resource management and scheduling, while supporting rapid instance recreation and rollback in the event of a failure.
2. Secure isolated computing environment for the Clevi-X MDLM model
Clevi-X MDLM (M Double LM - Massive Multimodal Language Model) is a next-generation multimodal AI model designed to support real-world computing tasks by learning how humans use computing environments, as well as their methods, structures and procedures.
A lightweight container environment based on CLK (Clevi Linux Kernel) is designed to make virus intrusion and malicious access, as well as any operation beyond authorised permissions, impossible.
Even in cloud computing environments accessed by multiple anonymous users for various purposes, SVCE(Safe Virtual Compute Environment) completely isolates each session and task, ensuring security against intrusions or attack attempts by malicious users.
As all tasks, including internet access, computation, code execution, result verification and file input/output, are performed in a secure isolated space, the reliability of the model's output data and its ability to execute tasks are ensured.
Provides a secure environment against various threats, including hacking, data leakage, attacks using the model and attacks targeting the model.
