Using HPE GreenLake and Red Hat OpenShift to deploy SAS Viya
Here in this blog, we will learn how to deploy SAS Viya using HPE Greenlake and Red Hat Open Shift.
Based on a contemporary, containerized microservices-based architecture, SAS Viya is a high-performance AI and analytics platform that facilitates quick and flexible data-driven decision-making. SAS Viya enables collaboration and scaling while operationalizing discovery and decision-making for users of all skill levels and organizational levels, from executives to business analysts to data scientists.
The solution architecture developed by SAS for SAS Viya deployment on-premises in partnership with Red Hat and Hewlett Packard Enterprise (HPE) is covered in this blog. It makes use of Red Hat OpenShift, the industry-leading hybrid cloud application platform driven by Kubernetes for the deployment and scaling of containerized applications, and the HPE GreenLake edge-to-cloud platform, which offers the cloud experience where the data resides.
While the public cloud facilitates the deployment and consumption of IT and IT services, a number of SAS customers are encountering challenges when attempting to shift their entire SAS 9 landscapes to the public cloud. The primary causes are:
- Their data’s hybrid and distributed nature
- Concerns about security and performance
- Diminished ability to control their data
- Not enough time or money to prepare, test, and execute
- Policies pertaining to consolidation, compliance, and data protection
- Excessively expensive egress
Many customers are now considering an on-premises or hybrid cloud approach in light of these factors. As a result, there is mounting evidence to support SAS Viya’s on-premises implementation.
Summary of the solution
SAS has developed a Red Hat OpenShift solution architecture for SAS Viya that can be used in hybrid and on-premises cloud environments. With this solution, HPE offers a platform that runs OpenShift on the most recent HPE GreenLake Compute and Storage offerings, including HPE GreenLake for Block Storage based on HPE Alletra Storage MP and HPE ProLiant Gen11 servers. Red Hat OpenShift Virtualization is used to completely virtualize this offering.
Options for services
Customers can choose from a variety of solution management service options with this solution, including partner-managed, HPE-managed, and self-managed options. HPE is capable of handling OpenShift’s installation, configuration, and daily management along with the underlying HPE GreenLake infrastructure. SAS, a SAS partner, or the customer may be in charge of managing SAS Viya.
Parts of the solution
OpenShift Red Hat
Hat in Red OpenShift reduces the friction associated with developing, modernizing, deploying, running, and managing applications by bringing together reliable and tested services. Whether using edge architecture, hybrid cloud, on-premises, or public cloud, OpenShift offers a consistent user experience.
Installing SAS Viya on OpenShift offers a unified, adaptable platform to handle a range of business requirements, from a robust, enterprise-ready Kubernetes orchestrator to an all-inclusive platform for developing cloud-native applications that can be used as a fully managed cloud service or left to self-manage.
Furthermore, Red Hat Advanced Cluster Management for Kubernetes offers numerous benefits to the deployment as a whole, such as enhanced security via security access controls, governance and policy enforcement compliance, and easier deployment and management.
The solution, which consists of OpenShift-hosted control planes (HCP) with OpenShift Virtualization running on bare metal HPE ProLiant servers, is depicted in Figure 2 as its component parts. This solution supports the installation of multiple tenant OpenShift clusters using worker nodes operating on virtual machines (VMs) thanks to HCP and OpenShift Virtualization.
OpenShift Virtualization by Red Hat
You can run and manage virtual machine workloads in addition to container workloads with OpenShift Virtualization. As part of the HPE Solution for SAS Viya on OpenShift, OpenShift Virtualization enables you to deploy SAS Viya on a tenant OpenShift cluster operating on virtual machines.
Nodes in the control plane
A virtual or bare metal computer is called a node in a Kubernetes cluster. The Kubernetes services required to manage the Kubernetes cluster, including the Kubernetes controller manager, etc, scheduler, and API server, are hosted on control plane nodes. Additional OpenShift services, such as the OpenShift OAuth API server, OpenShift controller manager, OpenShift OAuth server, and OpenShift API server, are operated by the control plane nodes.
The management cluster is deployed as a compact three-node cluster to provide a highly available OpenShift control plane that identifies the three control plane nodes as schedulable, allowing them to double as worker nodes for the execution of applications. For the worker nodes in the tenant guest OpenShift cluster(s), additional bare metal worker nodes in the management cluster facilitate the operation of virtual machines (VMs) via OpenShift Virtualization. In this solution, the tenant guest clusters are supported by the OpenShift management cluster, which functions as infrastructure.
Based on the open-source project HyperShift, HCP is used by the management host cluster to aggregate the control planes for several tenant guest clusters in hosted mode. Tenant cluster control planes are hosted by the management cluster and operate as multiple pods within distinct namespaces for each guest cluster. The benefits of integrating HCP with OpenShift include the following:
- Improved security posture as a result of separating the data plane and control plane
- Reduce expenses by using fewer nodes to run multiple control planes
- Control plane startup times are accelerated since they are made up of pods that are launched on OpenShift
Employee nodes
The SAS Viya application container pods are hosted by worker nodes. Many worker nodes can be supported in OpenShift environments; however, node redundancy necessitates at least two worker nodes. SAS developed a number of reference configurations, also known as t-shirt sizes, in collaboration with HPE and Red Hat to give clients a place to start. The number and size of worker nodes as a whole depend on the application resource requirements plus extra capacity needed to schedule node maintenance tasks like cluster updates.
Workloads are assigned to the appropriate worker nodes within the tenant cluster using SAS Viya workload classes and node pools. By applying predefined Kubernetes node labels and node taints within multiple node pool definitions, the placement of SAS workload classes is made possible. While using node pools to place SAS Viya workloads is not required, it does have many advantages, including the ability to customize the cluster topology to the needs of the workload by defining distinct hardware configurations for nodes with GPU cards, extra storage, and other features. Additionally, SAS Viya workload classes like SAS Compute can enable cluster node autoscaling of a node pool.
HPE GreenLake Data Center and Processing
When viewed through the HPE GreenLake perspective, this solution consists of HPE ProLiant Gen 11 servers outfitted with 4th generation Intel Xeon Scalable processors and HPE GreenLake for Block Storage, which is built upon HPE Alletra Storage MP.
Built on top of HPE Alletra Storage MP, HPE GreenLake for Block Storage is an all-NVME, scale-out, disaggregated block storage solution with the ability to scale capacity and performance separately. Viya’s HPE Container Storage Interface (CSI) driver exposes the storage to the OpenShift layer for dynamic volume provisioning.
SAS Viya
There are three different license packages available for SAS Viya: SAS Viya, SAS Viya Advanced, and SAS Viya Enterprise. Each package includes small, medium, and large t-shirt sizes. To accommodate the nine distinct SAS Viya license t-shirts, Red Hat and HPE have developed five distinct hardware configurations in this solution.
Python, Java, R, and Lua are just a few of the programming languages other than SAS that SAS Viya can integrate code written in and access data from almost anywhere. It provides a number of low-code or no-code UI-driven tools for managing data, creating models, and implementing insights. Natural language processing (NLP), computer vision, neural networks, generative AI, machine learning (ML), deep learning, intelligent decisions, advanced analytics, and more are among the fields in which SAS Viya provides solutions. Both public cloud and on-premises support are available for it. An OpenShift on-premises solution is provided in this article.
Conclude
For your AI and analytics use cases, SAS Viya on HPE GreenLake and Red Hat OpenShift offer a scalable and effective on-premises solution. Utilizing the most recent HPE ProLiant Gen11 servers, this solution makes use of HPE GreenLake for block storage based on HPE Alletra Storage MP and OpenShift and OpenShift Virtualization technologies. With the HPE GreenLake edge-to-cloud platform, you can manage, configure, and keep an eye on your whole IT environment from a single, cloud-native pane of glass. Instead of having to relocate your data to a specific location in order to take advantage of these advantages, HPE GreenLake gives you the flexibility to keep your data wherever it is and extend the cloud experience and cloud-like consumption to it.