DKube

Cloud Native Deep Learning Solution

Software defined network implementation

Features

Composable
Simplicity

Integrated, intuitive data scientist & operator workflow enables users to focus on problem-solving rather than plumbing

Solution will work out-of-the-box

Data Scientist working on models within 4 hours, including installation

Familiarity with tools offers fast learning ramp

Customizable
Based on Open Platforms

Enables best-in-class components, including TensorFlow, PyTorch & Jupyter

Ability to rapidly add future frameworks & algorithms

Components integrated and validated together for robust operation

Heterogenous
Mobility & Scalability

Migration from on-prem to cloud, and between cloud providers

Linear performance improvement by adding nodes to cluster

No cloud or vendor lock-in

Deep Learning
Collaboration

Support for multi-tenancy

Ability to securely partition system into separate groups

Share models, datasets, and resources

Scalable
On Demand GPU Allocation

GPUs are seamlessly allocated to users from a flexible pool, both within a server and between servers

API Driven
Efficient scale-up or scale-out

Support for heterogeneous clusters with device access across servers

High bandwidth, low latency communication through RDMA

API Driven
Security

Enterprise-grade solution

Authentication of users

Secure authorization for models, datasets, and resources

API Driven
Customizable

One Convergence can tailor the solution to your specific needs

Complete Data Scientist Workflow
  • Integrates deep learning components for experimentation, training & inference
  • Dynamic administration of users, groups and resources
  • Intuitive workflow management
  • Support for the latest frameworks and algorithms
  • Simple, flexible version control
  • Collaboration among users
Deep Learning
Deep Learning
API Driven
Secure Collaboration
  • Secure policy-based collaboration mechanism
  • Dynamic allocation of users to groups
  • Unauthorized user access prevented
  • Users in groups share models, resources, & data
  • Flexible, granular access permissions
  • Heterogeneous, distributed pooling of resources across servers
Deep Learning
Mobility & Scalability
  • Cloud Native solution
  • Large, diverse, & growing family of applications
  • Built on open platforms
  • Range of off-the-shelf systems
  • Choose the best balance of performance, flexibility, and cost
  • Easy migration from on-prem usage to the cloud, and between cloud providers
Mobility & Scalability
Flexible GUI
  • Intuitive operation
  • Charting options
  • Operator & Data Scientist dashboards
  • Accessible through APIs
  • Drag & drop management
Deep Learning
API Driven
Support & Services
  • Tiered levels of support
  • Integration with custom authentication systems
  • Tuning & optimization for target markets or specific platforms
  • Flexible integration of custom workflow stages in the pipeline
Support & Services