Release Notes

ZEBLOK COMPUTATIONAL AI-MICROCLOUD RELEASE NOTES

November 2024

Basic(v1.3.2)

1. KMS Implementation for Datasets in Basic Edition

• Improved Data Security: With Key Management Service (KMS) now integrated into the Basic SKU, dataset encryption and key management have been significantly enhanced.

• Seamless Data Access: Users can securely manage encryption keys and access datasets without compromising performance.

2. Inference HUB in Ai-MicroCloud®

• Centralized Model Management: Introducing Inference HUB, a dedicated space in AiM for hosting, managing, and deploying inference models.

• Enhanced Collaboration: Share, update, and deploy AI models with ease, ensuring faster time-to-market for your AI solutions.

• SDK Compatibility: Fully integrated with the updated SDK for smooth development and deployment workflows.

3. Bug Fix for Pipeline: Delete Feature and Persistent Logs

• UI Improvements: The delete feature in the Pipeline UI has been fixed, ensuring a smoother and more reliable user experience.

• Accurate Logs: Logs are now persistent, providing comprehensive insights into pipeline activities and debugging.

4. SDK Update to Support Inference HUB

• Streamlined Workflows: The SDK has been upgraded to seamlessly integrate with the new Inference HUB.

• Improved Compatibility: Simplifies interactions between datasets, models, and inference services, enabling smoother development cycles.

Enterprise (v1.3.1)

1. HostPort and HostPath in AI-Microservices

• Custom Port Mapping: Configure HostPort settings for streamlined service deployment across hosts.

• Flexible Storage: Leverage HostPath support to mount host directories directly into pods, simplifying data management.

2. Multiple Services for Multiple Ports

• Enhanced Deployment Flexibility: Now deploy multiple services, each with its own dedicated port mapping, for complex workloads.

• Improved Efficiency: Simplifies deployment processes while supporting diverse application needs.

3. Network Policies

• Strengthened Security: Implement network policies to regulate traffic between pods and safeguard workloads.

• Fine-Grained Controls: Customize ingress and egress rules to meet organizational security standards.

4. Root Privileges Access

• Elevated Access for Critical Operations: Enable root privileges for specific workloads requiring administrative-level permissions.

• Compliance with Governance: Carefully managed access ensures security without compromising functionality.

5. Object Storage Bucket Attachment for Microservices

• Seamless Integration: Onboarded microservices can now attach to object storage buckets effortlessly.

• Scalable Solutions: Simplify data management with unified and scalable storage options.

October 2024

Basic(v1.3.1)

Platform as a Service (PaaS) Enhancements

• PVC/Storage Graph Integration into Monitoring:

  • We’ve integrated Persistent Volume Claims (PVC) and storage usage into our monitoring graphs, giving you better visibility into storage resources and allocations.

• ZBL-CLI Release Management:

  • We’ve streamlined the release management process for ZBL-CLI, making version control and deployments more efficient, reducing manual steps and simplifying updates.

• Multiple Delete Option:

  • You can now delete multiple resources at once, significantly speeding up resource management tasks and improving efficiency.

• MINIO in High Availability (HA) for On-Premise Setup:

  • We’ve configured MINIO in HA for on-premise environments, providing increased fault tolerance and resilience, ensuring continuous uptime and data availability.

• URL Authentication for MS/WS (Environment Authentication):

  • We’ve added environment-based URL authentication for Microservices (MS) and Workstations (WS), offering enhanced security by ensuring access is limited to specific environments.

• NGINX Approach to Utilize Refresh Token (URL Authentication):

  • Our NGINX setup now utilizes refresh tokens for URL authentication, adding an additional layer of security by validating tokens before granting access to services.

• Notifications and Alerts on Resource Spiking and Threshold Reaching:

  • Stay informed with real-time notifications and alerts when resources spike or reach defined thresholds, enabling proactive resource management and faster response times.

• Multiple Services for Additional Ports:

  • When adding extra ports, the system now creates multiple services automatically, providing more flexibility in managing network traffic.

• Automatic Bucket Mounting on MS/WS for User-Onboarded Images (S3/Blob/MINIO):

  • We’ve enabled automatic bucket mounting for user-onboarded images across S3, Blob, and MINIO, streamlining cloud storage integration for Microservices and Workstations.

IaaS Enhancements

• Network Policies Implementation:

  • With the introduction of network policies, traffic between pods and services is now more controlled, enhancing overall security and enabling more granular control over network communications.

• Module Verification Script (Testing Only):

  • We’ve introduced a Module Verification Script for testing purposes, ensuring that newly developed features are thoroughly verified before full integration.

SaaS Enhancements

• SDK Validation Checks, Dataset, and Multi-File Upload Feature:

  • Our SDK now includes validation checks to ensure data integrity, along with the ability to manage datasets and upload multiple files simultaneously, offering greater flexibility in data handling.

Enterprise (v1.3.0)

Platform as a Service (PaaS) Enhancements

Notifications and Alerts:

• All user-related activities, as well as component activities such as MS (Microservice), WS (Workstation), AI-API, and Containerization, are now tracked in the Notification Tab.

• Errors during spawning, running, or deletion are logged and can be accessed easily by users.

• Notifications reset once all items are marked as read, with new notifications briefly popping up on the user’s workspace.

Microservice and Workstation Updates

• Edit Feature for Microservices and Workstations:

  • Users can now update spawned microservices and workstations, allowing modifications to attached resources, environment variables, and ports.

• Microservice and Workstation Sharing:

  • Microservices and workstations can now be shared within an organization, offering resource optimization by avoiding redundancy and enabling collaboration.

Stateful Set Integration

• Workstation:

  • All workstations have been converted into StatefulSets, offering persistent storage and stable network identity throughout their lifecycle.

• Microservices:

  • Users can now choose to deploy microservices as stateful components by selecting the relevant checkbox during the creation process.

• Replica Sets:

  • Replica sets ensure high availability and efficient storage management.

    • Scaling: Users can adjust the replica count to scale applications easily.

    • Fault Tolerance: Failed pods are automatically replaced, maintaining desired replica numbers.

Backend Enhancements

• Private Registry Support:

  • The issue with secret creation due to special characters has been resolved. Microservices from private registries can now be easily onboarded and spawned.

• URL Authentication:

  • Security has been enhanced to ensure that users can only access services within the MicroCloud. Unauthorized access is prevented, as users are redirected to the login page if they attempt to access services without valid credentials.

User Interface Enhancements

• Improved Containerization and Pipeline UI:

  • The deployment tab has been renamed "Containerization," encompassing all CaaS activities with detailed logs for users.

• Information Carousel at Spawning Sections:

  • A carousel at the top of spawning sections provides important updates or messages related to the activities in the section below.

• Form Field Examples and Hover Buttons:

  • Examples and tooltips in form fields guide users on correct input formats, reducing errors and improving usability.

IaaS v1.3.0 Enhancements

Improved Security:

• Stronger Encryption: Encryption standards have been upgraded to use stronger algorithms like AES-256, enhancing protection.

• Authentication: Enhanced with Elliptic Curve Cryptography (ECC) for secure key exchanges.

• Forward Secrecy: Perfect Forward Secrecy (PFS) is supported, ensuring that past communications remain secure even if a long-term key is compromised.