YellowDog’s core solutions can be used in isolation or combination for intelligent hybrid cloud scheduling and orchestration.
Securely accelerate processes across on-premise, private, and public cloud providers
Choose, provision, and control multiple sources of compute through a single API
Self-healing and predictive scheduling manages workload deadlines effectively and reduces costs
Under the hood.
The YellowDog Platform comes complete with its own scheduler. It can also be integrated with existing schedulers if required. If the scheduler within YellowDog is being used then, once configured, the YellowDog Platform splits the batch workloads into smaller segments before sending the job out to the worker nodes. This segmentation, and subsequent parallelisation where possible, of the processing tasks, is what accelerates the workload processing time.
Compute Orchestration automatically brings up servers in the public and private cloud providers depending on the requirements of the workload. Connectors to the leading public cloud providers is provided out of the box and YellowDog takes care of the heterogenous nature and the different ways that the different providers’ infrastructure works.
Once the job is complete and results securely returned, and there is no more need for compute, the YellowDog Platform gracefully brings the servers down so that costs are controlled.
User-defined criteria and workload requirements determine which pools of compute are available. Through an intuitive user interface, business rules are defined to ensure the best source of compute is used every time.
The criteria used could be one or many of:
• Cost of that compute to always ensure the lowest cost.
• Computing power or type of compute available (CPU or GPU, how many processing cores at any time).
• Geographical location of the servers to preserve data sovereignty and ensure compliance.
• Distance between processing nodes if network latency is a constraint.
Pools of compute can be made up of servers on-premise and in public or private clouds.
In shared, virtualised environments, the performance of virtual machines (VMs) varies depending on the load overall physical server. For example, where a single server has four VMs sharing its resources, and three of those VMs are busy with high compute load tasks, the fourth VM will suffer performance degradation. This is in contrast to the performance of a similar specification VM that is located on a server that is less busy.
In order to take account of this performance degradation (commonly referred to as “noisy neighbours”) and ensure that the right amount of processing power is available to deliver a workload to its deadline, the YellowDog Platform automatically adds or removes worker nodes to the allocated resource of a workload based on the actual processing power being made available at that point in time.
Forward Prediction Analytics predicts how long jobs will take to complete and the likelihood of failure of these jobs. The service employs Machine Learning and YellowDog proprietary algorithms to deliver these features.
The YellowDog Performance & Analytics Dashboard displays a performance snapshot with a customisable time frame. This functionality helps System Administrators, Managers, and Analysts understand how YellowDog is helping to improve utilisation of existing server infrastructure and accelerate processes. The Dashboard is securely accessed through a web browser. This data can be exported by customers for inclusion in documents and presentation slides.
Contact the YellowDog team to learn more or to request a demo.
You are seeing this because you are using a browser that is not supported. The YellowDog website is built using modern technology and standards. We recommend upgrading your browser with one of the following to properly view our website:Windows
Please note that this is not an exhaustive list of browsers. We also do not intend to recommend a particular manufacturer's browser over another's; only to suggest upgrading to a browser version that is compliant with current standards to give you the best and most secure browsing experience.