Today, data is the most valuable resource. Leading databases such as Microsoft SQL Server assist businesses in optimizing transactional performance and unleashing real-time insights from data, making it one of the most popular databases. With the growing popularity and adoption of containers and Kubernetes, especially for stateful applications, Microsoft is putting a great deal of effort into bringing Microsoft SQL Server 2019 to Linux containers. In fact, in a recent Diamanti survey of IT decision-makers, databases emerged as one of the top container use cases.
Challenges of Containerizing Microsoft SQL Server
Containers and Kubernetes provide database applications portability, unparalleled agility, and flexibility to support a DevOps approach of continuous integration and continuous delivery (CI/CD). However, these stateful applications exhibit specific key properties, which introduce inherent challenges left for the underlying platform to handle. These challenges are:
- SQL Server applications demand much greater I/O than standard applications and are more sensitive to latency variations. Since I/O is such an essential part of SQL Server performance, enterprises need to make sure their infrastructure is not the bottleneck
- These relational databases track inventories, process customer information and manage vast amounts of mission-critical e-commerce transactions, and much more. This mission-critical data demands SQL Server applications to run on platforms that offer state of the art data services to ensure security, high availability (HA), disaster recovery and data protection
- SQL Server applications work with massive amounts of data and require large disk space to store that data. As the amount of data and the number of SQL Server instances grow, the underlying platform needs to scale in a cost-effective way to minimize the total-cost-of-ownership (TCO)
- Enterprises are increasingly deploying Microsoft SQL Server and supporting applications in hybrid cloud environments. This means that the underlying platform needs to provide seamless migration and failover/failback of SQL Server databases across on-premises and public clouds
The aforementioned challenges leave even the most seasoned IT professionals, DevOps engineers, and platform architects wondering which Kubernetes platform would be the most optimal for running their I/O intensive Microsoft SQL Server databases across on-premises and public cloud environments.
Performance Comparison of SQL Server on Leading Kubernetes Platforms
To address these challenges, Diamanti commissioned McKnight Consulting Group, an independent consulting firm, to compare the performance of Microsoft SQL Server 2019 on the best possible configurations of the following Kubernetes platforms:
- Diamanti Enterprise Kubernetes Platform
- Amazon Web Services Elastic Kubernetes Service (AWS EKS)
- Azure Kubernetes Service (AKS)
This report provides insights and results to companies looking to invest in a cost-effective Kubernetes platform that delivers the highest performance while maintaining consistently low latency to run business-critical applications like Microsoft SQL Server. This report also helps enterprises identify the right platform supporting their hybrid cloud strategy, giving them the flexibility to deploy SQL Server to the most appropriate infrastructure driven by multiple factors such as cost, security, performance requirements, and geography.
Summary of Results
The benchmarking results are stunning – Diamanti is 4x faster and 10x cheaper while running Microsoft SQL Server compared to Azure with Azure Ultra disks and 6x faster and 13x cheaper compared to AWS Nitro with IO2 disks.
Built with Performance at its Core
Diamanti Kubernetes management platform differentiates itself from conventional architectures by leveraging a container-native storage and networking plane that offers high-performance NVMe storage and quality of service guarantees across performance tiers. For on-premises environments, the Diamanti platform offers Ultima offload, a pair of second-generation PCIe based I/O acceleration cards that offload networking and storage traffic, freeing up compute resources to power modern applications. This minimizes IO wait time to ~0, giving customers 3x-30x performance improvement on data-intensive applications while getting consistent sub-100 microsecond latency and 1 million IOPS per node.
Footprint Reduction and TCO Savings
These performance improvements directly translate into TCO savings as enterprises can drive greater resource utilization and reduce their overall footprint and operational costs.
Enterprise-grade Storage Services
Diamanti platform offers a full spectrum of data availability features such as snapshots, mirroring, replication, disaster recovery( DR), backup and restore to ensure that the enterprise deployments of Microsoft SQL databases are always online and healthy. Diamanti allows enterprises to migrate and seamlessly failover/failback(DR) applications across data centers and public clouds. Additionally, the platform provides software (volume encryption) and hardware-based (self-encrypting drives) security measures to protect against data theft and unintentional data loss.
To learn more about this validated design and testing results, download the Executive summary.
Stay tuned for a full report on these findings and access to the test results.