Hyperscale Next: Rack Integration
Download PDFWed, 04 Jan 2023 19:03:53 -0000
|Read Time: 0 minutes
Summary
The PowerEdge Product Group and Services, in partnership with Intel®, have purpose-built a program to directly support the growth and specific needs of Cloud Service Providers. This includes hosters, B2B Service Providers, and B2C Service Providers like you. We are offering the Hyperscale Next program to select pre-approved accounts. This Direct from Development (DfD) tech note summarizes how our rack integration service enables rapid time-to-value by delivering ready to deploy rack solutions.
Hyperscale Next
As the public cloud market accelerates business operations and functionality for customers, the largest hyperscalers dominate the cloud infrastructure market and often get favorable treatment from supplying vendors, including advanced access to new technologies, helping them market faster with differentiated services. Unfortunately, this makes it harder for other CSPs, hosters, and online service providers to provide differentiation and innovation. Most must wait until general availability to test new technologies. It can take them several months to take new technology from testing to full production, making it difficult to respond rapidly to evolving technologies and customer demands. The Dell Hyperscale Next program is designed to help our cloud customers react to evolving technologies and customer demands, expand their cloud services and applications, and provide differentiated performance for target workloads. The program includes early access to new technology, rack integration, factory server configuration, ProSupport One, Self-service support technology, and server performance tuning.
Rack integration
A single rack can comprise hundreds of individual components that need to be identified, ordered, assembled, and eventually tested. These components are acquired through various suppliers, with lead times ranging from days to months. In addition, components must be inspected, tracked, and stored until the rack is assembled. Given the complexity of acquiring the necessary components, a single part delay can wreak havoc on even the best planned schedules! Designing, integrating, and testing all components in a rack that meets your unique environment is a complex process. The process requires expertise, including but not limited to server and rack engineering, networking, software integration and automation, hardware validation, purchasing, transportation, and logistics. The great news is that the Hyperscale Next Rack Integration Service lets Dell experts manage the complexities of rack integration and deliver fully configured and tested rack solutions that can be deployed within hours of reaching your data center.
How does it work?
It starts with a Dell Project Manager. This person works to understand your unique requirements and is your single point of contact for the project. They scope the project, keep you updated, and work with the various Dell teams to develop a plan that works for you and your schedule. Let us take a closer look at the major steps.
Planning and design
Planning is critical to a successful and on-time rack integration project, and we analyze every detail. During this phase, we verify component specs, review the bill of materials, finalize the project requirements, and order materials. In addition, Dell has developed relationships with suppliers worldwide to ensure that we can get our customers what they need, when they need it.
Configuration collection
We create a rack specification document for your unique rack solution. This document includes the layout, servers, power distribution, networking, labels, and cabling. If your solution requires custom cables, we can work with our partners to make it happen. In addition, the document includes details on required firmware settings, firmware versions, and your software installation or configuration. Finally, this document is machine-readable, so our automated rack test suite can use it to validate that each component is installed and configured precisely to your specifications. Before proceeding, our project manager reviews the document with you for approval or updates.
Factory build
Dell has rack integration sites worldwide where we have perfected the science of rack integration. First, the materials are collected and securely stored until assembly. Then, after everything is in place, our integration team inspects the components and assembles them to your specification. Components that go into a rack are shipped in packaging to ensure that they reach their final destination intact. The packaging protects equipment but generates waste. The Dell rack integration service allows Dell to recycle and reuse packing material that may otherwise find its way into a landfill. As outlined in our FY21 ESG Progress Made Real Report, Dell has committed that by 2030, 100% of our packaging will be made from recycled or renewable material. Read more about what Dell is doing to advance sustainability here.
Server testing, imaging, and provisioning
Using the BIOS, iDRAC, and other component settings defined in the rack specification document as input, our automated rack integration framework configures settings on each server. This step saves our customers hours because they do not have to perform onsite manual provisioning or develop automation to do it themselves. After configuring the servers, our automation framework performs basic functional tests and checks server event logs for errors. In rare instances, a hardware failure occurs, and we have direct access to replacement parts to allow us to move forward quickly.
We also use the rack integration framework to install custom software or operating systems during server provisioning in a consistent and repeatable manner. Our software development team is also available to work with you to integrate other testing requirements, such as burn-in or load testing, to minimize or eliminate your onsite provisioning to further your rapid time to value.
Network, power, and serial testing
The automated rack integration framework collects the MAC data from all servers and switches and confirms correct network connectivity across the rack by comparing it to the rack specification document. We repeat the process for any serial devices. Also, the integration framework methodically cycles power on smart PDU sockets and confirms that connected devices power off when they should.
Final test and asset report
We consolidate the test results into a final test report, or “green sheet” after completing the build, testing, and provisioning process. This green sheet tells the team that a rack is ready to ship! Collecting asset information can be tedious. Therefore, our integration framework collects all asset info, such as network card MAC and Dell service tag data, during the integration process. We then combine the data into a carefully formatted report that our customers can programmatically read to update their asset information before rack delivery, saving even more time.
First article inspection
We conduct final rack inspections and a first article inspection where we select a random rack and manually confirm every requested configuration change. We invite our customers to participate in person for the first article inspection, or we record videos and photos of the inspection and share them before shipping.
Shipping and installation
The racks are loaded into trucks, securely strapped using extensively tested standards, and sent to their final destination. Upon arrival, Dell can physically install the racks in a data center or hand off to your team. We also offer onsite validation, where we can rerun server, network, serial, and power testing.
Let’s get started!
Using the Dell Rack Integration service gives our customers rapid time to value by taking the complexity out of designing, building, and deploying enterprise rack solutions. Our experts use years of partner relationships, experience, and best practices to make the process painless and efficient. Work with your Dell account team for more detailed information or email to hyperscale.next@dell.com.
References
Related Documents
Hyperscale Next: Performance Tuning
Wed, 04 Jan 2023 18:41:31 -0000
|Read Time: 0 minutes
Summary
The PowerEdge Product Group and Services, in partnership with Intel, have purpose-built a program to directly support the growth and specific needs of Cloud Service Providers. This includes Hosters, B2B Service Providers, and B2C Service Providers like you. We are offering the Hyperscale Next program to select pre-approved accounts. This Direct from Development (DfD) tech note summarizes how our performance engineering team can enable rapid time-to-value through data-driven performance optimization testing at no-cost.
Hyperscale Next
As the public cloud market accelerates business operations and functionality for customers, the largest hyperscalers dominate the cloud infrastructure market and often get favorable treatment from supplying vendors, including advanced access to new technologies, helping them market faster with differentiated services. Unfortunately, this makes it harder for other CSPs, Hosters, and online service providers to provide differentiation and innovation. Most must wait until general availability to test new technologies. It can take them several months to take new technology from testing to full production, making it difficult to respond rapidly to evolving technologies and customer demands.
The Dell Hyperscale Next program is designed to help our cloud customers react to evolving technologies and customer demands, expand their cloud services and applications, and provide differentiated performance for target workloads — with speeds unmatched by other Tier 1 providers. The program includes early access to new technology, rack integration, factory server configuration, ProSupport One, Self-service support technology, and server performance tuning.
A Workload Optimized Server
Dell Technologies understands that time-to-value is critical when making IT purchasing decisions. Our no-cost Hyperscale Next Performance Tuning Program aims to give select customers a head start when evaluating and adopting new enterprise technologies. Optimizing a server for a particular workload can be challenging, time-consuming, and costly. Our performance tuning program allows you to work with our world-class performance engineers to configure and optimize a PowerEdge server for your unique workload. This direct engagement lets our engineers understand your requirements and perform iterative testing in our lab to make data-driven recommendations for optimal workload performance.
Figure 1: Dell PowerEdge R650xs
The Process
IT decision-makers must answer a few questions when mapping a workload or use case to a server.
- Server: Which server meets my workload and use-case requirements?
- CPU: Which CPU SKU delivers the best price/performance? More cores or higher frequency?
- Storage/Memory: What are my capacity and performance requirements?
- BIOS and Firmware: Which settings deliver the best performance for my workload?
- Networking: What is my networking bandwidth requirement per CPU?
- Power: What are my power constraints?
- Cost: What is my budget?
Our team works with you to identify the appropriate server, components, and server settings for the workload.
For example, when running a transactional database, selecting a low core count and high-frequency CPU SKU can deliver high performance and be more cost-effective due to licensing costs. Likewise, a memory configuration with the optimal DIMM quantity, capacity, and speed increases performance and minimizes TCO. Here’s how it works.
- Dell performance engineers meet with you to understand your workload requirements
- Our engineers map your workload to synthetic benchmarks that act as proxies for the workload
- Our engineers perform iterative benchmarking in our lab to study the effects of various BIOS, firmware, OS settings, and system settings on workload performance
- Finally, our engineers analyze the data to identify the hardware configuration and server settings that deliver optimal performance for your workload
Performance Tuning
Our performance engineers use popular synthetic benchmarks as proxies for your workload to evaluate performance and identify server settings that optimize performance. Some of the example workloads we use are shown in Figure 2.
Benchmark | What It Tests |
MLC and Stream | Memory Bandwidth and Latency |
FIO | Storage IOPS and Throughput |
SPEC CPU 2017 (SPECint) | CPU Integer Performance |
HPL (Linpack) | CPU Floating Point Performance |
HammerDB | SQL, MySQL database performance |
SPECjbb 2015 | Java Server Performance |
iPerf | Network Bandwidth |
Figure 2: Synthetic Benchmarks List
Our engineers select the appropriate benchmarks in coordination with your team. Then, using the benchmarks, we perform iterative testing in a Dell performance lab to analyze the effects of specific server settings and hardware configurations on a benchmark. This data-driven approach with engineers specializing in PowerEdge system performance allows Dell to identify the optimal system configuration for a given workload and provide guidance that delivers rapid time to-value for our cloud customers.
Performance Troubleshooting
The Hyperscale Next Program includes direct access to our system performance engineers. When evaluating a new server or optimizing your current fleet, our team helps you quickly understand and resolve issues limiting expected performance in your workload or benchmark. This service is another example of how we are helping our customers achieve rapid time-to-value when they choose Dell PowerEdge servers.
Early Access Program/s
Through the Hyperscale Next Program, select customers receive server engineering samples to start their evaluation before a platform RTS. This early access program enables rapid time-to-value by allowing you to evaluate and optimize for your environment before a product is publicly available. With the help of our Hyperscale Next team, you get a head start on adopting the latest technologies. In addition, the performance tuning service is available for platforms in the early access program and currently shipping products.
Let’s get started!
The Hyperscale Next Program performance tuning service is available to select customers at no additional cost. This program helps our customers enable and optimize the latest technologies at the pace their business demands. Work with your Dell account team for more detailed information or email hyperscale.next@dell.com.
VMware Cloud Foundation on Dell Hyperconverged Infrastructure (Dell EMC vSAN)
Fri, 13 Jan 2023 10:00:33 -0000
|Read Time: 0 minutes
Summary
Dell EMC vSAN Ready Nodes are preconfigured, tested and certified hyperconverged infrastructure to run VMware Cloud Foundation. Each Ready Node includes the right amount of CPU, memory, network I/O controllers, within a vSAN hyperconverged infrastructure cluster for a private cloud solution.
Rather than starting from the scratch of deploying varied virtualization tools for compute, network, storage, etc., vSAN Ready Nodes helps start from a preconfigured layer to deploy VMware Cloud Foundation, which helps customers provision infrastructure automatically and reliably through workload domains. These Ready Nodes help reduce human error by reducing the time taken to choose and validate the right cloud infrastructure, while greatly accelerating the time required to provision ready to use infrastructure from weeks to just few hours.
Introduction to the Technology
VMware Cloud Foundation is a unified software defined data center solution that bundles compute, storage, network virtualization and cloud monitoring and management platform. VMware Cloud Foundation on Dell EMC AMD vSAN Ready Nodes platform can be deployed either on- premises or hosted with a service provider as a private cloud or run as a service within or integrate to, a public cloud.
VMware Cloud Foundation Architecture
Cloud Foundation can be deployed as two types of architecture model - standard and consolidated. A standard architecture model includes a dedicated management domain, with dedicated infrastructure created during deployment phase, whereas a consolidated architecture is deployed on a smaller infrastructure footprint of six or fewer hosts, and both the management and workload domains run in the management domain.
VMware Cloud Foundation extends the benefits of compute, storage and network virtualization to legacy data centers and their cumbersome management of multiple siloes of infrastructure and simplifies the provisioning, orchestration, monitoring and management of infrastructure resources with an integrated software defined data center platform that is the foundation for both private and public cloud environments.
Infrastructure for hosting VMware Cloud Foundation as Standard Architecture Model
Dell EMC PowerEdge hyperconverged (VMware vSAN) infrastructure are one of the building blocks in the Dell EMC solution portfolio, wherein we can deploy the VMware Cloud Foundation solution stack. A hyperconverged infrastructure (vSAN Ready Node) improves the efficiency and management of complex data center infrastructure by leveraging the capabilities that a vSAN Ready Nodes provide as the basic building block for the Cloud.
he following are the minimum hardware required for the Standard Architecture.
For Management Domain.
The Management domain hosts the infrastructure workloads and requires a minimum of 4-hosts.
For Virtual Infrastructure Workload Domain.
The Virtual Infrastructure Workload Domains host the tenant workloads and consists of a minimum of one cluster of 4 hosts or more. Hosts within a single cluster must be homogeneous.
The below table lists the infrastructure required to host the VMware Cloud Foundation as a Standard Architecture Model.
Description | Configuration | Qty | Notes |
Management Cluster (4-vSAN Ready Nodes) | 4 | 4-Node Management Cluster | |
| 1S 8C (for All Flash) | 2 | All Flash nodes within a cluster |
1S 8C (for Hybrid) | 1 | Nodes with a combination of flash and spindle disks within a cluster | |
| 192 GB | 1 |
|
| 16GB Boot Device | 1 | Local Media |
NVMe or SSD | 1 | Caching Tier | |
SSDs or HDDs | 2 | Capacity Tier | |
| 10 GbE or higher | 2 | (IOVP Certified) |
| 1 GbE (Optional) | 1 |
|
| |||
VI Workload Domain (4-Compatible vSAN Ready Nodes-cluster minimum) | 4 | Choose from the VCG list of validated vSAN RN for User Workload | |
|
|
|
As per user-workload Sizing |
|
|
| |
|
|
| |
|
|
| |
|
|
| |
|
Table-1. Hardware Configuration – For Standard Architecture
The below table lists the AMD Processor based servers validated for vSAN Ready Nodes.
PowerEdge vSAN RN | No. Of Sockets | No. Of Cores | AMD CPU Platform |
| 1S | 32/64 |
AMD EPYC 7502, AMD EPYC 7742 |
| 2S | 32/64 | |
| 1S | 32/64 | |
| 2S | 32/64 | |
| 2S | 32/64 |
Table-2. List of AMD platform-based PowerEdge vSAN Ready Nodes
In Conclusion
Dell PowerEdge AMD based vSAN Ready Nodes are pre-configured, validated and certified to reduce deployment challenges, reduce time and increase efficiency, and in-turn helps ease the setting up of a VMware private cloud.
The availability of a validated vSAN Ready nodes of different configurations help customers scale strategically and ensures that critical IT resources are configured as per standards for maintaining productivity and competitive cloud advantage.