Elastic and Hybrid Execution
On-Prem or Cloud, Scalable, Flexible
At KNIME, we strive to produce software that is flexible, providing you a variety of deployment options. That is why we created KNIME Executors. A KNIME Executor is the software that executes workflows for the KNIME Server. With the KNIME Server Large edition, KNIME Executors are able to run in their own virtual machines. This opens up a world of possibilities with KNIME Server Large.
What are those possibilities for your enterprise?
- Meet your computing capacity needs by scaling KNIME Executors
- Providing a compute environment to meet all configuration needs
- Flexible deployment options to mix and match private datacenter and cloud resources
An enterprise needs the flexibility to cope with scaling out to handle greater workloads and the elasticity to deal with workloads dynamically, adding more resources when the workload increases and shrinking down to remove unneeded resources when demand gets less.
KNIME Executors provide enterprises with the options to be able to respond efficiently to constantly changing demands. Combined with cloud services, Executors give the added flexibility of autoscaling. They can also be specialized for certain purposes e.g. access GPUs for faster training of deep learning models, large memory executors for massive datasets, executors with licenses for third party software, and more. By defining resources when launching the executors, you can choose which executor type a specific workflow should be executed on.
On KNIME Server Small and Medium, KNIME Executors are embedded. They co-exist on the same physical or virtual machine as KNIME Server, sharing resources.
Elastic Scalability, Cloud
KNIME Cloud Offerings - AWS and Azure
Meet compute demands elastically and automatically with KNIME’s cloud offerings: AWS and Azure.
- KNIME Server Large bring your own license (BYOL)
- Multiple Executors that can be used by KNIME Server Large
- Pay as you go (PAYG) offering supports elastic scaling
- BYOL uses cores from your existing Server license
KNIME Executor Autoscaling
KNIME Executors are designed to work with the autoscaling capabilities provided by cloud platforms: New executors are started and shut down as required, saving money on both instance and license costs.
- Cloud platform autoscaling capabilities
- Elastic scaling to meet demand - scaling out to meet needed capacity, scaling in to save costs
- Supported on AWS Marketplace in PAYG licensing model
Flexible Hybrid Deployments to Meet Computing Needs
Enterprises need the flexibility to meet periodic demand peaks and keep to a budget. At times, specialty hardware (e.g. GPUs) is required to run operations. These needs can be met dynamically with a hybrid setup - a mix of the enterprise’s data center and cloud deployments.
Additional benefits of a hybrid setup: mix in cloud deployment to augment your current KNIME Server license temporarily until you can determine what capacity you really need. This alleviates you from the need to know your capacity at the time of purchasing a KNIME Server.
Install KNIME Server on premise and start up Executors in the cloud (on demand PAYG or BYOL) as needed.
Supporting Executors in the Enterprise with Executor Groups
Do you need to segregate execution resources logically, e.g. by department? The introduction of Executor Groups enhances the already existing KNIME Executors and simplifies the scaling process. Match users/groups to Executor Groups; partition compute resources by groups or department; partitioning is managed by your Server administrators.
Each group contains one or more executors and workflows are routed to the correct group, guaranteeing availability of execution resources.