Your Browser is Out of Date

Nytro.ai uses technology that works best in other browsers.
For a full experience use one of the browsers below
Reference Architecture Guide—Ready Solutions for Microsoft SQL: Design for Dell EMC XtremIO
Executive summary
Reference architecture overview
Audience
We value your feedback
Physical architecture Logical architecture Hardware and software components Design considerations Validation and use cases Conclusion References Appendix A: Best practice guidance Appendix B: Configuring vSphere Appendix C: Creating and mapping storage to the VM

Reference architecture overview

Home > Workload Solutions > SQL Server > Guides > Reference Architecture Guide—Ready Solutions for Microsoft SQL: Design for Dell EMC XtremIO > Reference architecture overview

Reference architecture overview

  • Today’s IT organizations are under pressure to achieve cloud-like elasticity, scalability, and ease of provisioning while lowering the total cost of ownership—a challenging goal, particularly with the complexity of databases. To achieve the anticipated business outcomes requires a clear statement of associated success metrics that often must be negotiated with the executive sponsor. After the success metrics have been defined, the next challenge for the IT organization becomes selecting the technologies that will meet or exceed the success metrics. 

    Traditionally, selecting a database infrastructure was a long, exhaustive process with a wide variety of complexities that took months to resolve. What if there was a solution where all the components were validated and tested together using real SQL Server workloads? Prevalidation engineering means that the database solution is a proven platform, thus removing most of the complexity and time that is associated with manual integration work. Testing such a solution is more complex. A simple online transaction processing (OLTP) workload test only shows how the system performs if it is solely dedicated to one database. This approach is great for showcasing strong performance but falls short when you want to measure the performance of a multiple-database ecosystem. 

    A better test is to show how the database solution scales while supporting several SQL Server databases. Scalability is the capability of the database solution to support existing workloads with the potential to accommodate more databases for future growth. The traditional challenge with servers and storage has been growth of the database ecosystem, which has among the most resource-demanding and latency-sensitive applications. For example, as more databases are added to existing infrastructure, processor and storage contention can affect performance. Scalability in the cloud era means greater growth potential while performance remains consistent. Tradeoffs, such as sacrificing application responsiveness due to the growth of the database ecosystem, affect performance and cost of ownership. Today’s IT organizations are looking for database solutions that offer far greater scalability and performance to meet their success metrics.

    The new Ready Solutions for Microsoft SQL Server reference architecture has been validated with Dell EMC PowerEdge servers and Dell EMC XtremIO X2 all-flash storage. In addition to validating the solution with SQL Server, the Dell EMC labs pushed the boundaries of scalability testing by running 16 virtualized databases in parallel—8 on Windows Server 2016 VMs and 8 on Red Hat Enterprise Linux (RHEL) VMs. Key test findings include:

    • The PowerEdge R840 servers demonstrated strong scalability. The database load on each of the two PowerEdge servers meant oversubscription of CPUs to the virtual machines. Each server had significant unused processor resources while delivering on performance. 
    • The XtremIO X2 array delivered sub-500-microsecond latencies while supporting 275,000-plus IOPS with 72 flash drives. The achievable IOPS per the XtremIO X2 specification sheet is 220,000 IOPS. We found no tradeoff between IOPS and latency during our tests on the XtremIO X2 array.
    • XtremIO X2 inline data reduction technology reduced the size of a 1 TB SQL Server database to 239 GB, for a data reduction ratio of 3.52 to 1. 
    • The reference architecture delivered substantial consolidation savings:
      • One PowerEdge R840 server supported eight virtualized SQL Server databases.
      • XtremIO X2 inline data reduction savings enabled greater consolidation on all-flash storage.

    This guide provides a detailed overview of the test findings, including a review of performance differences between SQL Server running on Windows Server and RHEL. 

    This reference architecture offers a great degree of sizing flexibility to meet business requirements. You can start with a minimal configuration that can grow incrementally or with larger configurations to support hundreds of databases. Having been validated with SQL Server, the architecture enables accurate sizing and faster time-to-value.