Can Your Database Keep Up?
HyperFlex systems provide you the performance you need
If your organization is like most, databases like Microsoft SQL Server are used throughout your operation. Maybe you rely on these databases for online transaction processing (OLTP), data warehousing, or batch processing. Perhaps you use business intelligence, online analytical processing (OLAP), or report generation tools. Most likely, you use a mix of these systems. Whatever the workloads, your databases need to perform if they are going to be able to keep your applications fed with data.
Cisco HyperFlex™ systems are a great platform for Microsoft SQL Server, Oracle, and SAP databases. Our all-flash configurations allow you to run your database and applications on one platform and get predictable performance every time. Read the case study on how CorpFlex used HyperFlex to deployed Microsoft and Oracle databases and ERP applications, reducing their storage footprint and TCO.
A distributed architecture for faster results
It starts with the HyperFlex Data Platform where all of the storage devices are combined into a single distributed, multitier, object-based data store that allows all cluster resources to participate in I/O responsiveness. Data is dynamically distributed across the data store and, as you scale the cluster, every new and existing component contributes processing power and storage capacity.
We designed the data platform around a log-structured file system that uses a caching layer to accelerate read requests and write responses.
- Fast writes: Incoming data is striped across the number of nodes that you define to meet your data availability requirements. When data is added or updated, the log-structured approach simply appends a new block and updates the metadata, requiring little use of the server’s processors.
- Fast responses: When data is moved from cache to disk, the data is deduplicated and compressed. This process occurs after the write operation is acknowledged, so there is no performance penalty for these operations.
- Fast reads: The most frequently used data is stored in the caching layer to accelerate read operations.
Faster I/O for faster results
Let’s take a look at networking in Cisco HyperFlex systems. Why? Because your databases and applications need massive amounts of east-west traffic bandwidth and low latency. With integrated fabric interconnects, you get high-bandwidth, low-latency unified fabric connectivity that carries all production IP traffic, hyperconvergence-layer traffic, and management traffic over a single set of cables. Every connection in the cluster is treated as its own microsegment, with the same level of security as if it were supported with a separate physical link, making the integrated network more secure.
Consistent network performance
The system is designed so that all traffic, even from different blade server chassis, reaches any other node in the cluster with only a single network hop. No other vendor can achieve this result because they build switching into the blade chassis—switching that adds latency. Our latency is deterministic, so you get consistent network performance for the data platform, and you don’t have to worry about network constraints on workload placement. This single-hop architecture accelerates east-west traffic so that the cluster—and your databases and applications—perform better.
High data availability
The system stripes and replicates data across nodes based on policies that you set to meet your data availability requirements. You can even replicate data to local or remote clusters for backup or disaster-recovery purposes. How reliable is it? In the unlikely event of a failure, systems configured with five nodes or more can keep running even if all drives fail on two nodes simultaneously. And, the cluster can nurse itself back to health. In the event of a drive or node failure, lost data is regenerated using the redundancy mechanisms built into the system. When complete, the system can sustain further failures. That’s a deployment you can count on.
Our distributed architecture allows every virtual machine (VM) to use the storage I/O operations per second (IOPS) and the capacity of the entire cluster, regardless of the physical location of the VM. Why does this matter? Microsoft SQL Server VMs frequently need higher performance to handle bursts of application or user activity. When VMs can access all of the resources in the cluster, performance skyrockets.
The Enterprise Strategy Group tested HyperFlex hybrid and all-flash systems using HCIBench, a tool that tests the performance of hyperconverged clusters. Baselining and iterative testing results showed that HyperFlex hybrid systems provides:
- 3x higher VM density
- 3x reduction in read/write latency
- 7:1 reduction in IOPS variability
We didn’t just handle more I/O. We handled it faster. While the other platforms delivered average response times ranging from 6.61 ms to 44.45 ms, HyperFlex responded in 2.46 ms. You can read the full report here.
Get started with HyperFlex
On top of all that, HyperFlex systems are easy to deploy, manage, and scale. With our HyperFlex sizing tools, you can determine which configuration is the right starting point of your workload. You can also take the risk and guesswork out of deployment with our Cisco® Validated Design guide for deploying Microsoft SQL Server Database on Cisco HyperFlex all-flash systems.