unleash the business value hidden in your data silos · 2019-09-17 · unleash the business value...

18
Unleash the Business Value Hidden in Your Data Silos Deploying SAP Data Hub on the SUSE CaaS Platform and Intel-based Dell EMC Infrastructure ABSTRACT SAP Data Hub provides a comprehensive solution for organizations that need to orchestrate and govern data stored in distributed silos. It enables data analytics and business intelligence teams to manage data in disparate systems through an intuitive “single pane of glass.” This unified view can be one of the keys to transforming data into services that help your organization differentiate your business and create new lines of revenue. In this white paper, we outline a proven path to SAP Data Hub based on the SUSE Containers as a Service (CaaS) Platform running on Dell EMC infrastructure with Intel ® Xeon ® processors. This paper covers example use cases for this solution and key requirements for the software and hardware architectures to support the SAP Data Hub environment. The intended audience for this paper includes professionals focused on IT operations, data analytics and business intelligence. DELL EMC WHITE PAPER

Upload: others

Post on 22-May-2020

9 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: Unleash the Business Value Hidden in Your Data Silos · 2019-09-17 · Unleash the Business Value Hidden in Your Data Silos Deploying SAP Data Hub on the SUSE CaaS Platform and Intel-based

Unleash the Business Value Hidden in Your Data SilosDeploying SAP Data Hub on the SUSE CaaS Platform and Intel-based Dell EMC Infrastructure

ABSTRACTSAP Data Hub provides a comprehensive solution for organizations that need to orchestrate and govern data stored in distributed silos. It enables data analytics and business intelligence teams to manage data in disparate systems through an intuitive “single pane of glass.” This unified view can be one of the keys to transforming data into services that help your organization differentiate your business and create new lines of revenue.

In this white paper, we outline a proven path to SAP Data Hub based on the SUSE Containers as a Service (CaaS) Platform running on Dell EMC infrastructure with Intel® Xeon® processors. This paper covers example use cases for this solution and key requirements for the software and hardware architectures to support the SAP Data Hub environment. The intended audience for this paper includes professionals focused on IT operations, data analytics and business intelligence.

DELL EMC WHITE PAPER

Page 2: Unleash the Business Value Hidden in Your Data Silos · 2019-09-17 · Unleash the Business Value Hidden in Your Data Silos Deploying SAP Data Hub on the SUSE CaaS Platform and Intel-based

DELL EMC WHITE PAPER

The information in this publication is provided “as is.” Dell Inc. makes no representations or warranties of any kind with respect to the information in this publication, and specifically disclaims implied warranties of merchantability or fitness for a particular purpose.

Use, copying and distribution of any software described in this publication requires an applicable software license.

Copyright © 2019 Dell Inc. or its subsidiaries. All Rights Reserved. Dell, Dell Technologies, EMC and other trademarks are trademarks of Dell Inc. or its subsidiaries. Intel, the Intel logo and Xeon are trademarks of Intel Corporation in the U.S. and/or other countries. SAP and other SAP products and services mentioned herein as well as their respective logos are trademarks or registered trademarks of SAP SE in Germany and other countries. Other trademarks may be the property of their respective owners.

Dell Technologies believes the information in this document is accurate as of its publication date. The information is subject to change without notice.

Published in the USA 9/19.

TABLE OF CONTENTS

INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

THE BUSINESS PROBLEM . . . . . . . . . . . . . . . . . . . . . . . 1 The business value of SAP Data Hub. . . . . . . . . . . . . . . . . . 2

EXAMPLE USE CASES . . . . . . . . . . . . . . . . . . . . . . . . . 2 Fraud detection. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 Manufacturing equipment maintenance . . . . . . . . . . . . . . . . . 3 Customer affinity recommendations . . . . . . . . . . . . . . . . . . 3

REQUIREMENTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 Software architecture . . . . . . . . . . . . . . . . . . . . . . . . . . 4

SAP Data Hub . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4SAP Vora Distributed Database . . . . . . . . . . . . . . . . . . . 5Persistent Database . . . . . . . . . . . . . . . . . . . . . . . . . 5Container Registry . . . . . . . . . . . . . . . . . . . . . . . . . . 5Optional Hadoop cluster . . . . . . . . . . . . . . . . . . . . . . . 5SUSE CaaS Platform. . . . . . . . . . . . . . . . . . . . . . . . . 5Administration Node . . . . . . . . . . . . . . . . . . . . . . . . . 6Kubernetes Master Nodes . . . . . . . . . . . . . . . . . . . . . . 7Kubernetes Worker Nodes . . . . . . . . . . . . . . . . . . . . . . 7Optional SUSE Cloud Application Platform . . . . . . . . . . . . . . 7Storage architecture . . . . . . . . . . . . . . . . . . . . . . . . . 7Dynamically provisioned storage volumes . . . . . . . . . . . . . . 8Software and systems management . . . . . . . . . . . . . . . . . 8

Hardware architecture . . . . . . . . . . . . . . . . . . . . . . . . . 9Compute . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .10Storage. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .10Network . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .10

KEY TAKEAWAYS . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11

BETTER TOGETHER: DELL EMC, INTEL, SUSE AND SAP . . . . . 11

LEARN MORE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11

APPENDICES AND REFERENCES . . . . . . . . . . . . . . . . . .12Links to web-based resources . . . . . . . . . . . . . . . . . . . . .12Dell Technologies products . . . . . . . . . . . . . . . . . . . . . . .12Bill of Materials – Networking . . . . . . . . . . . . . . . . . . . . . .14Bill of Materials – Kubernetes Cluster . . . . . . . . . . . . . . . . .15Bill of Materials – Storage Cluster . . . . . . . . . . . . . . . . . . .16

Page 3: Unleash the Business Value Hidden in Your Data Silos · 2019-09-17 · Unleash the Business Value Hidden in Your Data Silos Deploying SAP Data Hub on the SUSE CaaS Platform and Intel-based

1 DELL EMC WHITE PAPER

INTRODUCTIONEnterprise data is exploding, creating both a challenge and an opportunity. Companies are discovering ways to transform their data into services that help to differentiate the business and create new lines of revenue. Unfortunately, managing and fully utilizing the information stored in data silos (e.g., cloud databases, Hadoop clusters, social media feeds) has become incredibly complex due to requirements for security, governance and specialized training.

To help your organization address these challenges, SAP Data Hub provides a GUI-based, business-wide view of a broad array of data systems, databases and assets, enabling your data analytics and business intelligence teams to manage your entire data landscape through an intuitive “single pane of glass.”

In what is likely to become a trend, SAP uses open source software to focus on its core competencies, making the company a leader in mission-critical enterprise applications. This is the promise of the SAP Intelligent Enterprise. Supporting SAP with innovation, SUSE has been the market leader for SAP applications for more than 20 years. SUSE is the trusted and preferred open source platform for SAP customers who want to unlock data intelligence, drive innovation and run with the best.

As an example of this open source mandate, SAP Data Hub is deployed on a Kubernetes-compatible container platform. SUSE Containers as a Service (CaaS) Platform enables your organization to extend your SUSE Enterprise Linux for SAP environment to container-based application delivery.

Ultimately, competing and innovating with data requires end-to-end optimization throughout the entire data lifecycle — from data ingest at the edge to staging and preparing data for analytics to achieving actionable insights from AI. Throughout every stage of the data lifecycle, Intel’s data-centric platform, optimized for your software environments and libraries, provides highly performant data solutions.

THE BUSINESS PROBLEMToday’s business leaders are under increasing pressure to drive their businesses with data-driven decisions. This expectation presents a particular challenge for those executives who strive to bring together the right combination of disparate data sources to unlock new value for their businesses.

Figure 1 . SAP Data Hub data pipeline funnel

DATA PIPELINE DATA PIPELINESAP Data Hub

IoT Data Social Media Google Cloud Platform Microsoft Azure Salesforce Hadoop Amazon Web Services Third-party Data On-Premises Data

Machine Learning Business IntelligenceVisualization ERP CRM

Page 4: Unleash the Business Value Hidden in Your Data Silos · 2019-09-17 · Unleash the Business Value Hidden in Your Data Silos Deploying SAP Data Hub on the SUSE CaaS Platform and Intel-based

2 DELL EMC WHITE PAPER

This difficulty is compounded by the very nature of how data is collected and stored, which results in independent data silos that provide no easy way to make critical associations across them. These data silos might be stored geographically close together or far apart. Some might be built with on-premises resources and some housed in one or more public clouds. Valuable data is often found in structured and unstructured databases, Hadoop data lakes, data warehouses and even in text files. Gaining new insights into potential customers and business opportunities could involve nearly all of the data silos a company has available to it.

What’s more, finding a business application to perform the task isn’t even the hardest part. Businesses need to ensure that their data-analysis software investment can meet the scale of their current and future application and data landscape, as well as enforce data governance. Equally important is the resiliency and scalability of the underlying infrastructure. Experienced leaders know that enterprise-grade software is a poor investment unless it is built on enterprise-grade infrastructure.

THE BUSINESS VALUE OF SAP DATA HUBSAP Data Hub is a containerized solution designed to be deployed on enterprise-grade Kubernetes clusters, such as SUSE CaaS Platform. For more than 17 years, SAP has developed its software on SUSE® Linux Enterprise Server (SLES) and SUSE solutions, such as SUSE CaaS Platform. (Note: See SAP Note 2693555 for certified systems.)

SAP Data Hub is built on a next-generation data-aggregation model that does away with the need for expensive data warehouses. Instead, SAP Data Hub allows for data extraction and formatting to be done on the platform where the data resides. This is in contrast to the current practice of using cumbersome, single-use extract, load, transform (ELT) operations to populate data warehouses. SAP Data Hub provides formatted, refined and cleansed data from multiple sources directly to the data consumers.

SAP Data Hub leverages data pipelines, which are built from reusable application components. Data pipelines are computational models that are executed natively on the data source. They define what data should be gathered from which sources and how that data should be formatted at the source. Pipelines also specify the refinements and cleansing that each stream of data should go through to make it compatible with the other data streams in the pipeline. Finally, the data pipelines identify which consumer(s) the collated data should be sent to. Because SAP Data Hub does not need to persist data, it eliminates the need for expensive, scale-limiting data warehouses.

Data pipelines can be created through a graphical user interface to leverage existing data sources, such as SAP HANA®, SAP Vora®, Apache Spark and Apache Hadoop, as well as all major open and closed source OLTP, OLAP and NoSQL databases.

Before implementing a data-analytics solution, consider the specific problem you are working to solve. Below are some use cases for SAP Data Hub that can help you zero in on the type of solution you are pursuing.

EXAMPLE USE CASESThis section outlines potential use cases for SAP Data Hub built on SUSE Containers as a Service Platform. In general, SAP Data Hub excels in pulling information from multiple types of internal and external data resources to enable insight into very complex analytical problems. The use of machine learning and big data analytics platforms (SAP, Hadoop, MapR, Cloudera, etc.) requires access to large pools of unstructured data in a highly automated, systematic and secure way.

FRAUD DETECTIONCredit card fraud has become an epidemic, with losses in the billions of dollars. Financial institutions need the ability to create profiles that alert them to probable fraud on large volumes of transactions. The more information they can cross-reference, the more accurate their

Page 5: Unleash the Business Value Hidden in Your Data Silos · 2019-09-17 · Unleash the Business Value Hidden in Your Data Silos Deploying SAP Data Hub on the SUSE CaaS Platform and Intel-based

3 DELL EMC WHITE PAPER

models will become. SAP Data Hub can pull in transactional data from ERP systems, credit reporting bureaus, email from a Hadoop cluster, social media data and “Dark Web” databases, enabling data scientists to build very precise detection methodologies.

MANUFACTURING EQUIPMENT MAINTENANCEGlobal manufacturers rely on the uptime of their equipment to meet product delivery targets. Unscheduled maintenance or equipment failure can result in lost profits, poor quality and unmet commitments. Conversely, over-scheduling maintenance activities also impacts cost and output. Manufacturers were early adopters of Internet of Things (IoT) technology for the real-time monitoring of equipment sensors (temperature, vibration, humidity, motor loading, etc.) to gain a better understanding of the state of their environments. What if predictive models and machine learning (ML) could be used to optimize maintenance scheduling?

SAP Data Hub can be used to orchestrate the end-to-end data flow needed to feed an ML platform to predict impending outages and then schedule corrective maintenance before any disruptions occur. In addition, using platforms with Intel® architecture, manufacturers can now easily advance AI, increase performance, use machine vision for defect detection and quality inspection, consolidate workloads, enhance security, and more.

CUSTOMER AFFINITY RECOMMENDATIONSE-commerce sites routinely use various data sources to recommend additional purchases or fine-tune searches to more relevant items. Early attempts were based solely on purchasing behavior at an individual retailer. However, state-of-the-art e-commerce now requires data input from email, social media, browser search data, clickstream data and credit card reporting sites. SAP Data Hub enables you to easily build this pipeline to feed real-time recommendations into an active session on a purchasing website. This information can greatly increase the revenue per transaction metric that is critical to success.

REQUIREMENTSAs your IT organization evaluates solutions to manage data growth and migration challenges, here are some key requirements to consider:

Requirement Type Details

Existing Data Stores

Access data from a variety of data sources, including Hadoop data lakes, object stores, databases and data warehouses, both in the cloud and on-premises.

• Perform data transformations, data quality and data preparation processes.• Define data pipelines and streams.• Embed and productize scripts, programs and algorithms of the data scientist.• Productize open libraries or ML algorithms in one framework.

Distributed Data ProcessingDistribute computational tasks to the native environments where the data reside. Enable remote process scheduling for:

• SAP Business Warehouse process chains• SAP Data Services dataflows• SAP HANA Smart Data Integration FlowGraph

GovernanceEstablish and manage zones in a landscape with attached policies and services levels.Establish security and access control capabilities.

OrchestrationCreate workflows for operations and processes across the landscape, with monitoring and analysis capabilities.Execute end-to-end data processes, starting with the ingestion of data into the landscape (e.g., the data lake), including data processing, and leading up to the delivery or integration of the resulting data into enterprise processes and applications.

Data Ingestion and Processing Focus on data integration, cleansing, enrichment, masking and anonymization.

Data DiscoveryDevelop data profiles for big data sets, showing quality and comprehensive structure information.Establish the ability to crawl, discover and tag data elements. Expose discovered data for further usage.

Scalability Develop a scalable architecture, from small to big, test to production deployment.Deployment Enable easy deployment, using a proven-to-work combination of the components.Fault Tolerance Establish fault tolerance, so single-component errors will not lead to the whole system being unavailable.Ease of Management/Ops Reduce complexity for solution management.Physical Footprint Gain more value with a compact solution that works within your existing infrastructure models.Flexibility Use a flexible building block approach that allows sizing according to customer needs.Security Provide the means to secure customer infrastructure.High Performance Leverage best practices designed into the solution to help ensure the best performance results.

Page 6: Unleash the Business Value Hidden in Your Data Silos · 2019-09-17 · Unleash the Business Value Hidden in Your Data Silos Deploying SAP Data Hub on the SUSE CaaS Platform and Intel-based

4 DELL EMC WHITE PAPER

With a solid understanding of your requirements, you can begin to design the solution. The following section outlines the key concepts in the software architecture of the SAP Data Hub reference configuration.

SOFTWARE ARCHITECTURE

SAP DATA HUBSAP Data Hub offers data management capabilities to help your organization manage your growing volume of data. This solution combines data governance, management of data pipelines and data integration, using a single visual interface and without the need for moving data into a central data warehouse. Figure 2 shows a high-level view of the architectural components designed to handle a wide range of enterprise applications scenarios. The optional Hadoop cluster can be used as the main software platform for handling the composition of application data.

Tenant applications and services are the core of SAP Data Hub. SAP Data Hub provides various tools for development and administration, as well as applications that are accessible through the SAP Data Hub application launchpad.

• SAP Data Hub Pipelines are the connectors between the various SAP Data Hub data sources. They provide reusable, configurable operations to process data from the various sources (including CSV files, web services APIs and SAP’s data stores) and can be flexibly designed.

• The SAP Data Hub Modeler allows for the creation and configuration of such pipelines through a graphical user interface.

• The metadata Explorer provides information about location, attributes, quality and sensitivity of data. With this information, you can make informed decisions about which datasets to publish and determine who has access to use or view information about the datasets.

• The Connection Management block enables connections to managed systems or external storage. Services such as Amazon S3, Google Cloud Services, Microsoft Azure (ADL, WASB), data services or Hadoop Distributed File System (HDFS) can be connected, as well as databases (Oracle, SAP HANA, SAP Vora) or business warehouses (SAP BW).

Figure 2 . Data Hub architecture

Docker Registry

Data Hub (Kubernetes Cluster)

Tenant Applications / Services

Vora Spark Extensions

HDFS / Spark

Pipeline MonitorFlow Agent

Metadata ExplorerDatabase Tools

Spark on KBSConnection Management

DiagnosticsSystem Management

Vora DatabaseTime Series Data

Persistent Database e.g., Metadata

Optional Hadoop Cluster

Page 7: Unleash the Business Value Hidden in Your Data Silos · 2019-09-17 · Unleash the Business Value Hidden in Your Data Silos Deploying SAP Data Hub on the SUSE CaaS Platform and Intel-based

5 DELL EMC WHITE PAPER

SAP VORA DISTRIBUTED DATABASESAP Vora is a horizontally scalable, distributed database that can store and process structured data, time-series data (i.e., IoT streams), graph data and semi-structured documents in-memory and/or on disk. SAP Vora is available only with SAP Data Hub, running in Kubernetes as a fully containerized application.

It can store analytics data in Kubernetes pods, as well as provide a bi-directional Spark2 interface between SAP Data Hub and an optionally co-located Hadoop cluster. PERSISTENT DATABASE

This database holds all the required persistent data required by SAP Data Hub (e.g., metadata). This instance is automatically installed, sized and maintained as part of the overall SAP Data Hub installation process. No special consideration is required.

CONTAINER REGISTRYSAP Data Hub requires a private and secure container registry to store and access its container images. This can be a publicly accessible site or a private collection of workload images. Although the private container registry is not part of the SUSE CaaS Platform, you can either:

• Build an on-premises instance using the Containers Module Add-on included with SUSE Linux Enterprise Server for SAP, along with the SUSE Portus (port.us.org) package; or:

• Deploy this as a container directly on SUSE CaaS Platform. Portus is an open source on-premises authorization service that enables users to administrate and secure their private container registries with fine-grained control.

OPTIONAL HADOOP CLUSTERAn optional Hadoop cluster can be built on dedicated nodes and co-located with SAP Data Hub. This associated Hadoop data lake can be used as a local computational/storage medium for SAP Data Hub original and uploaded content. The SAP Data Hub Vora Spark Extensions are used to interface with the Spark2 environment on the Hadoop cluster for processing and storing data.

When using this cluster, SAP Data Hub users can leverage the analytical strengths of SAP Vora to analyze and store data in HDFS through the SAP Data Hub Vora Spark Extension. In addition, SUSE has extensive experience deploying bare-metal and virtualized Hadoop clusters on SUSE Linux Enterprise Server. While this Hadoop cluster uses dedicated nodes, its HDFS storage is built on block storage from the SUSE Enterprise Storage cluster that also serves SAP Data Hub.

SUSE CAAS PLATFORMSUSE CaaS Platform is an integrated software platform that automates the tasks of building, managing and upgrading Kubernetes clusters. It combines the benefits of an enterprise-ready operating system with the agility of an orchestration platform for containerized applications, such as SAP Data Hub.

While there are several top-tier Kubernetes offerings in the market, SUSE CaaS Platform stands out for its ease of installation and configuration, DevOps integration (via SUSE Cloud Application Platform), and enterprise-level operability and scalability.

Page 8: Unleash the Business Value Hidden in Your Data Silos · 2019-09-17 · Unleash the Business Value Hidden in Your Data Silos Deploying SAP Data Hub on the SUSE CaaS Platform and Intel-based

6 DELL EMC WHITE PAPER

One of the biggest challenges for Kubernetes operators is matching the scalability of the node-level infrastructure with that of the overlaying container infrastructure. Inconsistently applied software changes, as well as node configuration drift, create ticking time bombs in production Kubernetes clusters.

SUSE CaaS Platform (Figure 3) resolves these problems with a combination of SUSE MicroOS as the container host operating system and Salt for configuration management. SUSE MicroOS is a mission-specific derivative of SUSE Linux Enterprise Server. While MicroOS leverages the same codebase and packages, its implementation helps ensure that software changes are applied atomically and within a snapshot-protected environment. The combination of MicroOS and Salt guarantees that all nodes in a cluster are always in a known and consistent state. The troubleshooting nightmares of discovering a single node with a partially failed configuration or software change are a thing of the past.

A SUSE CaaS Platform (Figure 4) consists of the following node types:

ADMINISTRATION NODEThe Administration Node of the SUSE CaaS Platform manages the deployment of the cluster and runs central services such as:

• Velum: Provides a web-UI dashboard used to administer the cluster

• Salt Master: Manages the configuration of the cluster nodes

• MariaDB Database: Stores Velum data and Salt master daemon events

• Dex Identity Service: Provides user authentication and a robust, role-based access control (RBAC) system

Figure 3 . SUSE CaaS Platform architecture

Container Container

Orchestration (Kubernetes) Services (e .g ., Deployment Dashboard)

Container Container Container

LoggingNetworking

Physical Infrastructure

SecurityPersistent Storage (local disk, NFS, SES) Registry

Automation (Salt + cloud-init) | Configuration and Management of each node

Container Runtime and Packaging | SUSE Linux Enterprise MicroOS (Container Host OS)

Kubernetes ClusterAdmin

S4048-ON

S3048-ON

AdminMasters MastersWorkers Workers

Storage Cluster

Access Network

Management Network

Page 9: Unleash the Business Value Hidden in Your Data Silos · 2019-09-17 · Unleash the Business Value Hidden in Your Data Silos Deploying SAP Data Hub on the SUSE CaaS Platform and Intel-based

7 DELL EMC WHITE PAPER

KUBERNETES MASTER NODESThe CaaS Platform Master Nodes maintain the Kubernetes control plane services. These services run as containers on the Master Nodes. While three or more Master Nodes (always an odd number) are required for high availability of the Kubernetes control plane, a single Master Node is acceptable for demonstration purposes.

KUBERNETES WORKER NODESThe CaaS Platform Kubernetes Worker Nodes run the SAP Data Hub application containers. SAP Data Hub requires a minimum of three Kubernetes Worker Nodes (four worker nodes for production). SUSE currently supports CaaS Platform clusters of up to 150 nodes. Additional Worker Nodes can be added to a Production CaaS Platform cluster non-disruptively.

(Note: SAP specifies that each worker node must have a least eight cores and 64 GB of main memory.)

OPTIONAL SUSE CLOUD APPLICATION PLATFORMSUSE Cloud Application Platform is a modern application delivery environment used to bring an advanced cloud-native DevOps experience to container-based infrastructure. SUSE’s implementation is based on the open source Project Eirini, which uses Kubernetes to orchestrate application containers while maintaining the Cloud Foundry user experience. This platform as a service (PaaS) environment is used by developers to streamline lifecycle management of traditional and cloud-native applications. Together, these technologies, based on Intel Technologies, accelerate innovation, improve IT responsiveness and help maximize return on investment.

STORAGE ARCHITECTUREThe storage layer of this solution leverages the software-defined storage capabilities of SUSE Enterprise Storage (SES). SES is a commercially supported distribution of the Ceph enterprise-grade, scale-out storage solution. SAP requires a certified solution for storage that supports Reliable Autonomic Distributed Object Store (RADOS) Block Devices as well as Dynamically Provisioned Volumes. (See SAP Note 2686169 for certified storage options.)

(Note: SAP Data Hub 2.x no longer supports the NFS protocol. See SAP Note 2712050.)

Figure 4 . SUSE CaaS Platform Node configuration

Dashboard

Admin Node

HA Proxy

Master Node

Worker Node

Worker Node

Master Node

Worker Node

Worker Node

Master Node

Worker Node

Worker Node

External Clients

SUSE Enterprise Storage

Kube

rnet

es

Clus

ter

Page 10: Unleash the Business Value Hidden in Your Data Silos · 2019-09-17 · Unleash the Business Value Hidden in Your Data Silos Deploying SAP Data Hub on the SUSE CaaS Platform and Intel-based

8 DELL EMC WHITE PAPER

Ceph is a scale-out, distributed object store that provides excellent performance, scalability and reliability. In most use cases, clients use Linux kernel libraries to read and write object and block data directly to/from a storage node in the SES cluster. SES also provides gateway options to support data access via iSCSI, NFS, S3 and Swift protocols.

The storage capacity of the SES solution can be expanded easily by integrating additional storage nodes into the cluster. Existing storage nodes will take care of redistributing the data to the newly added nodes without interrupting the availability of storage services to the clients.

SES provides a reliable, scalable storage layer for the complete solution, which supports:

• Dynamically provisioned block storage volumes to the pods running on SUSE CaaS Platform

• (Optionally) block storage volumes for the co-located Hadoop cluster nodes, if configured

• Object storage through an S3-API-compatible interface, for additional data storage and backups

DYNAMICALLY PROVISIONED STORAGE VOLUMESIn addition to providing block storage to the optional Hadoop cluster, a pod running on CaaS Platform can gain access to dynamically provisioned Kubernetes persistent volumes (PV) through Kubernetes persistent volume claims (PVCs). Persistent volumes are created as block devices in the supporting SES cluster. CaaS Platform uses PVCs to obtain dynamically provisioned persistent volumes through the software-defined storage mechanisms in SES. When a PVC is removed, the persistent volume and its associated block storage device in SES are automatically removed.

SOFTWARE AND SYSTEMS MANAGEMENTWhile SAP Data Hub doesn’t require an external SAP HANA instance in order to function, most users of this solution will attach to an existing HANA database to build their data pipelines. After assembling this combined data pipeline and writing to your HANA database, you can take advantage of SAP Advanced Analytics Processing capabilities, including machine learning/predictive analytics, spatial intelligence (location awareness) and streaming data processing.

The scale-out capabilities of SAP HANA support rapid data growth, but it is important to have a dependable method of updating your SAP HANA servers. SUSE Manager can mirror CaaS Platform installations and update packages to help enforce consistency across your organization. SUSE Manager can also analyze the container images in your private container registry as well as containers running on your SUSE CaaS Platform for known vulnerabilities, outstanding patches or pending package updates. SUSE Manager enables you to efficiently manage a set of Linux systems and keep them up to date.

An SAP HANA scale-out setup offers these benefits:

Reduce the complexity of managing SAP HANA environments .

• Ensure consistent management of SAP HANA and all other cluster systems.

• Manage your data environment across physical, virtual and cloud environments.

• Manage your channels effectively.

Create and manage development, QA and production channels .

• Add and manage third-party channels.

• Simplify compliance.

Page 11: Unleash the Business Value Hidden in Your Data Silos · 2019-09-17 · Unleash the Business Value Hidden in Your Data Silos Deploying SAP Data Hub on the SUSE CaaS Platform and Intel-based

9 DELL EMC WHITE PAPER

Audit the patch status for SAP HANA and subsystems .

• Track configuration changes and make sure all your administrators have the right authority for changes.

• Slash costs of ownership.

Automate system management tasks for SAP HANA and all other subsystems .

• Leverage a single, web-based interface to see the status of all your servers.

• Use your resources effectively.

HARDWARE ARCHITECTUREThis reference defines a private proof-of-concept cluster. Other guides will address cloud-based and production environments. The proof-of-concept cluster is a starting point for prototyping real-world applications that are meant to go into production. As such, a proof-of-concept cluster can easily be grown into a production environment. Further, a private deployment is useful for secure, in-house prototyping, but is not restricted from using cloud-based applications and data. SAP Data Hub can manage data from locations both behind the firewall and in the cloud.

The backbone of the application environment is the Kubernetes cluster, as implemented in the SUSE CaaS Platform software. As described in the software architecture section above, this cluster can be implemented on 1U racked boxes with two-socket processors.

The application environment reflects a key aspect of growth, from prototyping to production. In a PoC, there are typically only a few application containers, because developers are gaining experience building, deploying and managing code as containers on the new software environment. As such, there are fewer demands on the hardware for resources, particularly main memory. Memory can easily be expanded as the cluster needs to grow. Processor speed, on the other hand, is something to consider carefully. It is not as easy to swap processors, so choosing an Intel® Xeon® Gold or Intel® Xeon® Platinum Processor with many cores is a good idea.

It is recommended to start with a minimum two-socket Intel Xeon Scalable processor with eight to 12 cores per CPU. Memory can start at 64 GB. Local storage on the cluster is used primarily for the operating system and any temporary application data. A common choice is a RAID1 configuration of operating system disks, each with a minimum of 256 GB. SSD is the recommended choice, because it speeds operations and doesn’t add considerable cost to the cluster in this case.

Networking is another key performance area, so 10-GbE NICs are a minimum recommendation. You can start with one NIC, but that is a bare minimum; two or more NICs will be needed for specific network requirements, based on your application needs. In particular, high-bandwidth data pipelines will move a lot of data to the cluster storage and can easily saturate a single NIC. Consider mapping one high-speed NIC to those storage requirements.

As described in the SUSE CaaS Platform section above, a minimum of five nodes will be required: one administration node, one master node and three worker nodes. A SUSE YES-certified Dell EMC platform could be used for the physical nodes of this deployment, as long as the certification refers to the major version of the underlying SUSE operating system required by the SUSE CaaS Platform release.

Page 12: Unleash the Business Value Hidden in Your Data Silos · 2019-09-17 · Unleash the Business Value Hidden in Your Data Silos Deploying SAP Data Hub on the SUSE CaaS Platform and Intel-based

10 DELL EMC WHITE PAPER

One key benefit of this data analytics implementation is that Dell EMC servers with Intel Xeon Scalable processors can fulfill each of the resource node’s computational requirements and additional storage needs. A Dell EMC-recommended hardware infrastructure is listed in the appendices of this document. In addition, the appendices contain some respective component and resource sizing guidelines for each of the node roles.

COMPUTEThe following considerations for the system platforms should be emphasized:

• Ensure that all similar system devices are consistent and up to date with regard to BIOS/UEFI/device firmware versions, to reduce potential troubleshooting issues later.

• Reset the BIOS setup configuration to the default setting, in order to have a known baseline configuration for consistency.

• If possible, set up RAID1 mirroring on the storage controller across a pair of drives for the operating system installation.

STORAGEAs discussed in the storage architecture section above, the storage layer of this solution leverages the software-defined-storage capabilities provided by SUSE Enterprise Storage, which takes advantage of the performance and reliability of Intel processors.

SAP Data Hub and SUSE CaaS Platform are the base framework for a data analytics environment. The data analytics you execute will be defined by a set of application containers that run on the SUSE CaaS Platform. These containers will access data across your company’s infrastructure and may store derived results in SUSE Enterprise Storage. As you define the workflow of your data analytics applications, you will need to access data across many different storage systems in your enterprise. This access is beyond the scope of the architecture described herein, but it is important to understand that data access will be required for a wide range of disparate storage systems.

The SUSE Enterprise Storage cluster can be used to store both intermediary and final results from the data analytics pipeline. In other words, as your data analytics applications derive new data results, those are typically stored in SUSE Enterprise Storage. This enables the data analytics environment (SAP Data Hub, SUSE CaaS Platform and your data analytics applications) to be logically organized in one physical location.

NETWORKThe following considerations for networking should be emphasized:

• Configure 802.3ad for system port bonding in order to get the maximum performance of bonded network interfaces.

• Ensure that all similar switching devices are consistent and up to date with regard to firmware versions, to reduce potential troubleshooting issues later.

Page 13: Unleash the Business Value Hidden in Your Data Silos · 2019-09-17 · Unleash the Business Value Hidden in Your Data Silos Deploying SAP Data Hub on the SUSE CaaS Platform and Intel-based

11 DELL EMC WHITE PAPER

KEY TAKEAWAYSFor organizations that need to orchestrate and govern data stored in distributed silos, SAP Data Hub provides an ideal solution. It enables data analytics and business intelligence teams to manage data in disparate systems through an intuitive “single pane of glass.” This unified view can be one of the keys to transforming data into services that help your organization differentiate your business and create new lines of revenue.

The SUSE CaaS Platform running on Dell EMC infrastructure with Intel Xeon processors is an excellent environment for your SAP Data Hub implementation. This composable infrastructure based on Intel architecture enables your organization to define appropriate hardware from software descriptions. This means you can easily scale, adjust and customize your environment to fit your needs as you move from a proof of concept toward a production environment.

SUSE CaaS Platform is an enterprise Kubernetes container platform that provides a software foundation to support not only the SAP Data Hub software described in this reference architecture but also the data analytics applications you will build to ingest and manage your data. All of the software environments in this reference architecture are supported products and have been tested to work together on industry-standard x86-64 environments.

Ultimately, the SUSE CaaS Platform and Dell EMC infrastructure with Intel Xeon processors provide an ideal environment for your SAP Data Hub and related SAP solutions.

BETTER TOGETHER: DELL EMC, INTEL, SUSE AND SAP Together, Dell EMC, Intel and SUSE stand ready to work with your organization on the journey to SAP Data Hub. Our team can help your team achieve better outcomes by putting SAP systems to work to accelerate innovation and the intelligent enterprise.

The solutions for enterprises from Dell EMC, Intel, SUSE and SAP bring together all the hardware, operating and application software, services and expertise you need to design and deploy a comprehensive solution that spans a broad environment.

• Dell EMC provides a blueprint for industry-leading solutions that leverage proven infrastructure components and the powerful hardware, including Intel Xeon processors, required to drive real-time analytics.

• Intel processors empower Dell EMC edge, core and cloud solutions to run at peak capacity and enable world-leading benchmark performance.

• SUSE enables enterprises and other large organizations to deploy physical, virtual and cloud SAP workloads leveraging the SUSE Linux Enterprise Server operating system.

• SAP provides leading-edge solutions for the intelligent enterprise, in which data feeds intelligence, which in turn feeds process automation and innovation.

Ultimately, the combination of powerful hardware infrastructure components, sophisticated operating and application software, and a leading in-memory database enables your organization to leverage enterprise-class solutions that enhance your customers’ lives and help reduce your capital and operational costs.

LEARN MORETo learn more:• Contact your Dell EMC or SUSE account representative.• Contact [email protected] or visit suse.com/partners/alliance/dell/.

Page 14: Unleash the Business Value Hidden in Your Data Silos · 2019-09-17 · Unleash the Business Value Hidden in Your Data Silos Deploying SAP Data Hub on the SUSE CaaS Platform and Intel-based

12 DELL EMC WHITE PAPER

APPENDICES AND REFERENCESLINKS TO WEB-BASED RESOURCES

SAP DATA HUB• https://www.sap.com/products/data-hub.html

• SAP Note 2721708 — SAP Data Hub 2.4 Release Note on https://launchpad.support.sap.com

• "SAP Note 2686169 — Prerequisites for installing SAP Data Hub 2" on https://launchpad.support.sap.com

• Installation Guide for SAP Data Hub on http://help.sap.com

SUSE CAAS PLATFORM• https://www.suse.com/products/caas-platform

• Documentation — https://www.suse.com/documentation/suse-caasp-3

SUSE ENTERPRISE STORAGE• https://www.suse.com/products/suse-enterprise-storage

• Documentation — https://www.suse.com/documentation/suse-enterprise-storage-5/

DELL EMC NETWORK SWITCHES• S3048-ON — https://www.dell.com/en-us/work/shop/povw/networking-s-series-1gbe

• S4048T-ON — https://www.dell.com/en-us/work/shop/povw/networking-s-series-10gbe

DELL EMC POWEREDGE SERVERS • R640 Rack Server — https://www.dell.com/en-us/work/shop/povw/poweredge-r640

• R740xd Rack Server — https://www.dell.com/en-us/work/shop/povw/poweredge-r740xd

SUSE INTEL PARTNERSHIP• https://www.suse.com/partners/alliance/intel/

DELL TECHNOLOGIES PRODUCTS

DELL EMC NETWORK SWITCHES Dell EMC data center switching solutions are cost-effective and easy to deploy at any scale, from 1G to multi-rate 100G for optimum connectivity both within the rack or blade chassis and across the data center. The switching solutions also feature a choice of innovative Dell EMC and third-party software options to address virtually any enterprise or service provider use-case or environment.

For the physical switching layer, Dell EMC S-Series top-of-rack (ToR) open networking switches like the 10GbE S4048T-ON are cost-effective and easy to deploy at any scale for optimum connectivity both within the rack or chassis and across the data center fabric. The switches also feature a choice of innovative Dell EMC or third-party software options. To complete the networking solution, a Dell EMC Networking 1GbE S3048-ON open networking switch is used to handle management functions.

DELL EMC NETWORKING S3048-ONThe Dell Networking S3048-ON 1000BASE-T Top-of-Rack (ToR) switch is the industry’s first 1GbE enterprise switching platform to deliver both an industry hardened OS and support for open networking, providing freedom to run third-party operating systems (OS).

Page 15: Unleash the Business Value Hidden in Your Data Silos · 2019-09-17 · Unleash the Business Value Hidden in Your Data Silos Deploying SAP Data Hub on the SUSE CaaS Platform and Intel-based

13 DELL EMC WHITE PAPER

This open networking platform is built for high performance, software-defined data centers and provides the features to run traditional workloads and the flexibility to deploy new workloads such as Hadoop, SDS and big data. The S3048-ON switch offers the flexibility to run OS options optimized for diverse deployment needs on a common hardware platform and architecture.

The S3048-ON features a non-blocking switching architecture coupled with OS9.X software, delivering line-rate L2/L3 features for maximized network performance. The S3048-ON design provides (48) 1000BASE-T ports that support 10MbE/100MbE/1GbE and four 10GbE SFP+ uplinks. Each 10GbE interface can be used as uplinks to the network spine/core, as stack ports to connect up to six units in a stacked configuration, or a combination of both, depending on network architecture and uplink/stack bandwidth requirements.

The S3048-ON incorporates multiple architectural features that optimize data center network flexibility, efficiency and availability, including:

• I/O panel to PSU airflow or PSU to I/O panel airflow for hot/cold aisle environments

• Redundant, hot-swappable power supplies and fans with color coded touch points for ease of identification/ removal

• Dell ReadyRails for efficient installation of the switch into data center cabinets

The S3048-ON also supports the Dell Networking Embedded Open Automation Framework, which provides advanced network automation and virtualization capabilities for virtual data center environments. Embedded Open Automation Framework is a suite of network management apps that can be used together or independently to provide a network that is flexible, available and manageable while helping to reduce operational expenses.

DELL EMC NETWORKING S4048T-ONThe Dell EMC Networking S-Series S4048T-ON is a high-density 100M/1G/10G/40GbE top-of-rack (ToR) switch purpose-built for applications in high-performance data center and computing environments. Leveraging a non-blocking switching architecture, the S4048T-ON delivers line-rate L2 and L3 forwarding capacity within a conservative power budget. The compact S4048T-ON design provides industry-leading density of 48 dual-speed 1/10G BASE-T (RJ45) ports, as well as six 40GbE QSFP+ up-links to conserve valuable rack space and simplify the migration to 40Gbps in the data center core.

Each 40GbE QSFP+ up-link can also support four 10GbE (SFP+) ports with a breakout cable. In addition, the S4048T-ON incorporates multiple architectural features that optimize data center network flexibility, efficiency and availability, including I/O panel to PSU airflow or PSU to I/O panel airflow for hot/cold aisle environments, and redundant, hot-swappable power supplies and fans. S4048T-ON supports feature-rich Dell Networking OS, VLT, network virtualization features such as VRFlite, VXLAN Gateway and support for Dell Embedded Open Automation Framework.

In addition:

• The S4048T-ON is the only switch in the industry that supports traditional network-centric virtualization (VRF) and hypervisor-centric virtualization (VXLAN). The switch fully supports L2 VXLAN gateway function and has hardware support for L3 VXLAN routing.

• The S4048T-ON also supports the Dell EMC Networking Embedded Open Automation Framework, which provides enhanced network automation and virtualization capabilities for virtual data center environments.

• The Open Automation Framework comprises a suite of interrelated network management tools that can be used together or independently to provide a network that is flexible, available and manageable while helping to reduce operational expenses.

Page 16: Unleash the Business Value Hidden in Your Data Silos · 2019-09-17 · Unleash the Business Value Hidden in Your Data Silos Deploying SAP Data Hub on the SUSE CaaS Platform and Intel-based

14 DELL EMC WHITE PAPER

DELL EMC POWEREDGE R640 SERVERS The Dell EMC PowerEdge™ R640 server is the ideal dual-socket platform for dense scale-out data center computing and storage. Benefit from the flexibility of 2.5-inch or 3.5-inch drives, the performance of NVMe and embedded intelligence to help ensure optimized application performance in a secure platform. With embedded diagnostics and SupportAssist, the PowerEdge R640 server delivers maximum uptime in a worry-free environment.

Ideal workloads:

• Dense software-defined storage

• Service providers: application tier

• Dense private cloud

• Virtualization

• HPC

DELL EMC POWEREDGE R740XD SERVERS The Dell EMC PowerEdge™ R740xd server offers the benefits of scalable storage performance and dataset processing. This 2U, two-socket platform brings you scalability and performance to adapt to a variety of applications. Choose up to 24 NVMe drives, or a total of 32 x 2.5-inch or 18 x 3.5-inch drives. As you scale your deployments, scale your productivity with embedded intelligence and automation from iDRAC9 and the entire OpenManage portfolio designed to simplify the IT lifecycle from deployment to retirement.

Ideal workloads:

• Software-defined storage

• Big data server

• HPC

• Service providers: data tier

BILL OF MATERIALS – NETWORKING

Role Quantity Platform Configuration

Top-Of-Rack Network Switch Minimum 1 Dell EMC S3048-ON Connects up to 48 systems, add more as needed

Top-Of-Rack Network Switch Minimum 1 Dell EMC S4048T-ON 1 per rack of systems unless bonding is deployed, then 1 per linked interface.

Page 17: Unleash the Business Value Hidden in Your Data Silos · 2019-09-17 · Unleash the Business Value Hidden in Your Data Silos Deploying SAP Data Hub on the SUSE CaaS Platform and Intel-based

15 DELL EMC WHITE PAPER

BILL OF MATERIALS – KUBERNETES CLUSTER

Role Quantity Platform Configuration

Admin 1 Dell EMC PowerEdge R640 server

Chassis and power• 1 x 2.5-inch chassis with up to 8 hard drives and 3 PCIe slots• Riser configuration 2, 3 x 16 low profile• Dual hot-plug redundant power supply, 495 W• 2 x C13 to C14 PDU style, 12 AMP, 6.5 ft (2 m)

Processor and memory• 2 x Intel® Xeon® Gold 6138 processor, 2.0 GHz, 20C/40T, 27.5 MB L3 cache, 125 W• 6, 32 GB RDIMMs (192 GB)

Storage• Dell PowerEdge HBA330 Internal SAS HBA• 2 x 240 GB Intel SSD SATA DC S4600 Series, software RAID 1

(container runtime ephemeral storage)• Dell EMC BOSS controller card + 2 Dell/Intel 240 GB M.2 SATA SSD hard drive

SSDSCKJB240G7R, configured in hardware as RAID 1Connections and networking

• 1 x Intel X710 DP 10 GbE DA/SFP+, + i350 DP 1 GbE NDC• 1 x Intel XXV710 DP 25 GbE SFP28 PCIe adapter, low profile

Software and management• iDRAC 9 Enterprise

Master 3 Dell EMC PowerEdge R640 server

Chassis and power• 1 x 2.5-inch chassis with up to 8 hard drives and 3 PCIe slots• Riser configuration 2, 3 x 16 low profile• Dual hot-plug redundant power supply, 495 W• 2 x C13 to C14 PDU style, 12 AMP, 6.5 ft (2 m)

Processor and memory• 2 x Intel® Xeon® Gold 6138 processor, 2.0 GHz, 20C/40T, 27.5 MB L3 cache, 125 W• 6, 32 GB RDIMMs (192 GB)

Storage• Dell PowerEdge HBA330 Internal SAS HBA• 2 x 240 GB Intel SSD SATA DC S4600 Series, software RAID 1

(container runtime ephemeral storage)• Dell EMC BOSS controller card + 2 Dell/Intel 240 GB M.2 SATA SSD hard drive

SSDSCKJB240G7R, configured in hardware as RAID 1Connections and networking

• 1 x Intel X710 DP 10 GbE DA/SFP+, + i350 DP 1 GbE NDC• 1 x Intel XXV710 DP 25 GbE SFP28 PCIe adapter, low profile

Software and management• iDRAC 9 Enterprise

Worker 4 Dell EMC PowerEdge R640 server

Chassis and power• 1 x 2.5-inch chassis with up to 8 hard drives and 3 PCIe slots• Riser configuration 2, 3 x 16 low profile• Dual hot-plug redundant power supply, 495 W• 2 x C13 to C14 PDU style, 12 AMP, 6.5 ft (2 m)

Processor and memory• 2 x Intel Xeon Gold 6138 processor, 2.0 GHz, 20C/40T, 27.5 MB L3 cache, 125 W• 6, 32 GB RDIMMs (192 GB)

Storage• Dell PowerEdge HBA330 Internal SAS HBA• 2 x 240 GB Intel SSD SATA DC S4600 Series, software RAID 1

(container runtime ephemeral storage)• 2 x 960 GB Intel SSD SATA DC S4600 Series, software RAID 1

(OpenShift registry, logging and metrics)• Dell EMC BOSS controller card + 2 Dell/Intel 240 GB M.2 SATA SSD hard drive

SSDSCKJB240G7R, configured in hardware as RAID 1Connections and networking

• 1 x Intel X710 DP 10 GbE DA/SFP+, + i350 DP 1 GbE NDC• 1 x Intel XXV710 DP 25 GbE SFP28 PCIe adapter, low profile

Software and management• iDRAC 9 Enterprise

Page 18: Unleash the Business Value Hidden in Your Data Silos · 2019-09-17 · Unleash the Business Value Hidden in Your Data Silos Deploying SAP Data Hub on the SUSE CaaS Platform and Intel-based

16 DELL EMC WHITE PAPER

To learn more, visit DellEMC.com/sap.

BILL OF MATERIALS – STORAGE CLUSTER

Role Quantity Platform Configuration

Admin 1 Dell EMC PowerEdge R640 server

Chassis and power• 1 x 2.5-inch chassis with up to 8 hard drives and 3 PCIe slots• Riser configuration 2, 3 x 16 low profile• Dual hot-plug redundant power supply, 495 W• 2 x C13 to C14 PDU style, 12 AMP, 6.5 ft (2 m)

Processor and memory• 2 x Intel® Xeon® Gold 6138 processor, 2.0 GHz, 20C/40T, 27.5 MB L3 cache, 125 W• 6, 32 GB RDIMMs (192 GB)

Storage• Dell PowerEdge HBA330 Internal SAS HBA• 2 x 240 GB Intel SSD SATA DC S4600 Series, software RAID 1 (container runtime ephemeral

storage)• Dell EMC BOSS controller card + 2 Dell/Intel 240 GB M.2 SATA SSD hard drive

SSDSCKJB240G7R, configured in hardware as RAID 1Connections and networking

• 1 x Intel X710 DP 10 GbE DA/SFP+, + i350 DP 1 GbE NDC• 1 x Intel XXV710 DP 25 GbE SFP28 PCIe adapter, low profile

Software and management• iDRAC 9 Enterprise

Master 3 Dell EMC PowerEdge R640 server

Chassis and power• 1 x 2.5-inch chassis with up to 8 hard drives and 3 PCIe slots• Riser configuration 2, 3 x 16 low profile• Dual hot-plug redundant power supply, 495 W• 2 x C13 to C14 PDU style, 12 AMP, 6.5 ft (2 m)

Processor and memory• 2 x Intel® Xeon® Gold 6138 processor, 2.0 GHz, 20C/40T, 27.5 MB L3 cache, 125 W• 6, 32 GB RDIMMs (192 GB)

Storage• Dell PowerEdge HBA330 Internal SAS HBA• 2 x 240 GB Intel SSD SATA DC S4600 Series, software RAID 1 (container runtime ephemeral

storage)• Dell EMC BOSS controller card + 2 Dell/Intel 240 GB M.2 SATA SSD hard drive

SSDSCKJB240G7R, configured in hardware as RAID 1Connections and networking

• 1 x Intel X710 DP 10 GbE DA/SFP+, + i350 DP 1 GbE NDC• 1 x Intel XXV710 DP 25 GbE SFP28 PCIe adapter, low profile

Software and management• iDRAC 9 Enterprise

Worker 4 Dell EMC PowerEdge R740xd server

Chassis and power• 1 x R740xd chassis with up to 24 x 2.5-inch hard drives including max of 12 NVMe drives, 2 x

CPU configuration• Riser configuration 6, 5 x 8 and 3 x 16 slots• Dual hot-plug redundant power supply, 1100 W• 2 x C13 to C14 PDU style, 12 AMP, 6.5 ft (2 m)

Processor and memory• 2 x Intel® Xeon® Gold 6140 processor, 2.3 GHz, 18C/36T, 24.75 MB L3 cache, 140 W• 12 x 16 GB RDIMMs 2,666 MT/s (192 GB)

Storage• Dell PowerEdge HBA330 Internal SAS HBA• 2 x 960 GB Intel SSD SATA DC S4600 Series, software RAID 1 (container runtime ephemeral

storage)• 22 x 960 GB Intel SSD SATA DC S4600 Series (persistent storage)• Dell EMC BOSS controller card + 2 Dell/Intel 240 GB M.2 SATA SSD hard drive

SSDSCKJB240G7R, configured in hardware as RAID 1Connections and networking

• 1 x Intel X710 DP 10 GbE DA/SFP+, + i350 DP 1 GbE NDC• 1 x Intel XXV710 DP 25 GbE SFP28 PCIe adapter, low profile

Software and management• iDRAC 9 Enterprise