Kuiper

Deploy Everywhere. Optimize Everything.

Deploy workstations, headless workloads and more with unmatched control and flexibility.

How it Works

Slicing

Kuiper splits existing servers into "Slices" that can then be used to run different workloads. Everything from Workstations to Headless workloads can be run.

Dynamic Scheduling

Instead of statically assigning resources to workloads, Kuiper dynamically schedules workloads based on "requests" issued by Users and Applications. When a workload completes, the resources are returned to the pool and are ready to pick up new workloads.

This minimizes wasted resources, less technical support, and allows for a more flexible and dynamic environment.

GPU Slicing

With support for NVIDIA and AMD GPUs, Kuiper allows you to "compress" your workloads and increase workload density on all of your hardware.

Because Kuiper supports NVIDIA's GPU Operator, it allows for Time-Slicing. Time-Slicing functions without the need for a GRID license or vGPU license which provides additional savings. There are drawbacks to this method of GPU sharing, such as the inability to isolate GPU memory, but it is a great option for most workloads. Kuiper also supports MIG which does provide true isolation, but requires vGPU/GRID licenses to function. Kuiper allows you to mix and match these methods to get the best of both worlds.

Unmatched Hardware Compatibility

Your Hardware

Juno has a firm stance on "you bought it, you own it". We do not block hardware features behind a paywall, and we certainly don't limit its performance. We believe you know your hardware best, and you should be free to use it as you see fit.

Kuiper is built to run on any hardware. From the latest and greatest to the oldest and slowest. You can deploy kuiper to enterprise servers, workstations, laptops, and even Raspberry Pis. Kuiper was (and still is), built and run on micro Point of Sale devices ordered off of Amazon and Raspberry Pi 4s. You would be hard pressed to find a device that Kuiper cannot run on.

Optimization

Oversubscribed (Unlimited Resources)

Unlimited

Utilization

Usage: 0%

Undersubscribed

Wasted: 100%

Soft Oversubscription (35% Limited Resources with Burst Room)

35% Limit

Utilization

Usage: 0%

Undersubscribed

Wasted: 100%

* Oversubscription is supported on CPU and GPU workloads. RAM oversubscription is not supported due to upstream limitations with the built in OOM killer in Kubernetes. 3rd party solutions such as SDM (Software Defined Memory) can provide this functionality.

On-Prem Density

Kuiper can be used to reduce the number of on-prem servers needed to run your workloads by practicing "Density-On-The-Rack".

Environmental Benefits

Slicing inherently reduces the amount of servers needed to run your workloads. This means that you can reduce your carbon footprint by reducing the number of servers needed to run your workloads. Take those savings and reinvest in higher quality hardware to run even more dense workloads. Iterate faster and reduce your carbon footprint at the same time. This is a win-win for everyone.

Performance

More often than not, the hardware you have is underutilized. Kuiper allows you to "overcommit" your hardware and run more workloads. GPUs are a great example of this. For normal Workstation workloads, a single GPU can site relatively idle. But with Kuiper, you can "saturate" the GPU with multiple workloads, allowing you to squeeze every last drop of performance out of your hardware. While yes, this does come with some drawbacks, noisy neighbors, and performance degradation if misconfigured. But Kuiper gives you all the tools you need to make the right decisions for your workloads. Resource requests, limits, and priorities are all available to you. As mentioned before, this is your hardware, you know it best and we can't wait to see what you do with it.

Cloud Savings

Because Kuiper is built cloud native, it can apply all the same benefits from on-prem to the cloud while adding in the power of a Hyperscaler like AWS.

Kuiper leverages the power of Kubernetes to dynamically schedule servers on the fly to meet the needs of the requested workloads. This means that you only pay for the resources you need, when you need them. Once the server is up and running, Kuiper will then "Slice" the server into smaller virtual servers that can be used by multiple users at the same time. Allowing you to mix workstations, GPUs, and CPUs on the same server. This means that you can run multiple workloads on the same server, reducing the need for multiple servers and reducing costs.

g6.8xlarge

Kuiper Slices

g6.8xlarge = $2.01/hr

$2.01 / 4 slices = $0.50/hr per user

75 % reduction per user

Desktop Virtualization

Polaris Containerized Desktops

Kuiper by default deploys Juno's Polaris workstation container images. These are pre-installed with KasmVNC as a VDI solution. We support many different flavors of Linux. Such as Ubuntu, Fedora, Alma, Rocky and more.

Optimized for Performance

Polaris ships with a full XFCE desktop environment that is optimized for performance and usability. We do ship with a few default applications, but we encourage you to build on top of our base images to create your own custom images. This allows you to create a fully customized desktop experience for your use case.

VDI Support

While Polaris does ship with KasmVNC, we do support other VDI solutions such as NoMachine and RustDesk. There are many other VDI solutions that are available to use. We encourage you to try out different solutions to find the best fit for your use case. Even if we don't support it out of the box, you can always deploy your own VDI solution on top of Polaris and see if it suits your needs.

Microsoft Windows

As of now, we do not support Microsoft Windows as a base image. We are however working on a solution to support Windows in the near future. We will make sure to update the community when we have a solution for this. Stay tuned!

Pricing

By User

Features

Unlimited Workstations
GPU Slicing Support
CPU/RAM Allocation Tools

By Workstation

Features

Unlimited Users
GPU Slicing Support
CPU/RAM Allocation Tools

Cyber Operations

Features

Unlimited Workstations
GPU Slicing Support
CPU/RAM Allocation Tools
Minimized Attack Surface
Rapid Deployment/Teardown
Anonymous Login

All options come in Multi-Year, Yearly, or Monthly Subscriptions

Kuiper

Deploy Everywhere. Optimize Everything.

Deploy workstations, headless workloads and more with unmatched control and flexibility.

How it Works

Slicing

Dynamic Scheduling

GPU Slicing

Read more about the operators we support.

Unmatched Hardware Compatibility

Your Hardware

Optimization

Usage: 0%

Undersubscribed

Wasted: 100%

Usage: 0%

Undersubscribed

Wasted: 100%

On-Prem Density

Kuiper can be used to reduce the number of on-prem servers needed to run your workloads by practicing "Density-On-The-Rack".

Cloud Savings

Because Kuiper is built cloud native, it can apply all the same benefits from on-prem to the cloud while adding in the power of a Hyperscaler like AWS.

Kuiper Slices

g6.8xlarge = $2.01/hr

75 % reduction per user

Desktop Virtualization

Polaris Containerized Desktops

Kuiper by default deploys Juno's Polaris workstation container images. These are pre-installed with KasmVNC as a VDI solution. We support many different flavors of Linux. Such as Ubuntu, Fedora, Alma, Rocky and more.

Optimized for Performance

As of now, we do not support Microsoft Windows as a base image. We are however working on a solution to support Windows in the near future. We will make sure to update the community when we have a solution for this. Stay tuned!

KasmVNC

RustDesk

NoMachine

Pricing